{
  "schema_version": 1,
  "suite_questions": 120,
  "selected_records": 6000,
  "public_answer_records": 900,
  "configurations": [
    {
      "id": "gemma4-e4b/off",
      "suite_id": "tiny",
      "model": "Gemma 4 E4B",
      "mode": "Thinking off",
      "model_key": "google/gemma-4-e4b",
      "selected_variant": "google/gemma-4-e4b@q4_k_m",
      "quantization": "Q4_K_M",
      "adapter": "native-reasoning-v1",
      "generation": {
        "context_length": 32768,
        "max_output_tokens": 8192,
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 64,
        "min_p": 0,
        "repeat_penalty": 1
      },
      "thinking": "off",
      "reasoning_effort": "not separately specified",
      "generation_seed": null,
      "generation_seed_status": "not set; native API does not expose it",
      "runtime_version": "not verified",
      "exact_gpu_placement": "not verified",
      "source_id": "tiny/gemma-off",
      "batch": "tiny",
      "total_parameters": 7996156490,
      "active_parameters": null,
      "effective_parameters": 4500000000,
      "size_class": "medium",
      "parameter_count_basis": "Hugging Face safetensors.total for the pinned official artifact",
      "parameter_source_url": "https://huggingface.co/api/models/google/gemma-4-E4B-it/revision/ee0ef6023621cff504d758262d4e04895a5af4a2",
      "parameter_card_url": "https://huggingface.co/google/gemma-4-E4B-it/blob/ee0ef6023621cff504d758262d4e04895a5af4a2/README.md",
      "parameter_card_sha256": "b21e4f69614ccd77baa2f3797d05311040dee07b989cb9f0d25111aa4b605b2c",
      "parameter_checked_date": "2026-10-21",
      "parameter_notes": "Effective count is 4.5B; it is not an active-parameter measurement. Total includes embeddings and modality encoders."
    },
    {
      "id": "gemma4-e4b/on",
      "suite_id": "tiny",
      "model": "Gemma 4 E4B",
      "mode": "Thinking on",
      "model_key": "google/gemma-4-e4b",
      "selected_variant": "google/gemma-4-e4b@q4_k_m",
      "quantization": "Q4_K_M",
      "adapter": "native-reasoning-v1",
      "generation": {
        "context_length": 32768,
        "max_output_tokens": 8192,
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 64,
        "min_p": 0,
        "repeat_penalty": 1
      },
      "thinking": "on",
      "reasoning_effort": "not separately specified",
      "generation_seed": null,
      "generation_seed_status": "not set; native API does not expose it",
      "runtime_version": "not verified",
      "exact_gpu_placement": "not verified",
      "source_id": "tiny/gemma-on",
      "batch": "tiny",
      "total_parameters": 7996156490,
      "active_parameters": null,
      "effective_parameters": 4500000000,
      "size_class": "medium",
      "parameter_count_basis": "Hugging Face safetensors.total for the pinned official artifact",
      "parameter_source_url": "https://huggingface.co/api/models/google/gemma-4-E4B-it/revision/ee0ef6023621cff504d758262d4e04895a5af4a2",
      "parameter_card_url": "https://huggingface.co/google/gemma-4-E4B-it/blob/ee0ef6023621cff504d758262d4e04895a5af4a2/README.md",
      "parameter_card_sha256": "b21e4f69614ccd77baa2f3797d05311040dee07b989cb9f0d25111aa4b605b2c",
      "parameter_checked_date": "2026-10-21",
      "parameter_notes": "Effective count is 4.5B; it is not an active-parameter measurement. Total includes embeddings and modality encoders."
    },
    {
      "id": "lfm2.5-2.6b/native",
      "suite_id": "tiny",
      "model": "LFM2.5 2.6B",
      "mode": "Native thinking",
      "model_key": "lfm2.5-2.6b",
      "selected_variant": "not verified",
      "quantization": "Q4_K_M",
      "adapter": "native-default-thinking-v1",
      "generation": {
        "context_length": 32768,
        "max_output_tokens": 8192,
        "temperature": 0.1,
        "top_p": 1.0,
        "top_k": 50,
        "min_p": 0,
        "repeat_penalty": 1.1
      },
      "thinking": "on",
      "reasoning_effort": "not separately specified",
      "generation_seed": null,
      "generation_seed_status": "not set; native API does not expose it",
      "runtime_version": "not verified",
      "exact_gpu_placement": "not verified",
      "source_id": "tiny/lfm-on",
      "batch": "tiny",
      "total_parameters": 2697198592,
      "active_parameters": null,
      "effective_parameters": null,
      "size_class": "small",
      "parameter_count_basis": "Hugging Face safetensors.total for the pinned official artifact",
      "parameter_source_url": "https://huggingface.co/api/models/LiquidAI/LFM2.5-2.6B/revision/654f9463ce32b05d0429d76fe1f580b27d4c1ac0",
      "parameter_card_url": "https://huggingface.co/LiquidAI/LFM2.5-2.6B/blob/654f9463ce32b05d0429d76fe1f580b27d4c1ac0/README.md",
      "parameter_card_sha256": "003ceb63ec63a5b06b17f5e7b7a7d907695f037ad18720cabbcf8d8d68813f24",
      "parameter_checked_date": "2026-10-21",
      "parameter_notes": "Card rounds the total to 2.69B. Hybrid dense model; no separate active count is published."
    },
    {
      "id": "gemma4-e2b/off",
      "suite_id": "tiny",
      "model": "Gemma 4 E2B",
      "mode": "Thinking off",
      "model_key": "google/gemma-4-e2b",
      "selected_variant": "google/gemma-4-e2b@q4_k_m",
      "quantization": "Q4_K_M",
      "adapter": "native-reasoning-v1",
      "generation": {
        "context_length": 65536,
        "max_output_tokens": 32768,
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 64,
        "min_p": 0,
        "repeat_penalty": 1
      },
      "thinking": "off",
      "reasoning_effort": "not separately specified",
      "generation_seed": null,
      "generation_seed_status": "not set; native API does not expose it",
      "runtime_version": "not verified",
      "exact_gpu_placement": "not verified",
      "source_id": "tiny/gemmae2-off",
      "batch": "tiny",
      "total_parameters": 5123178051,
      "active_parameters": null,
      "effective_parameters": 2300000000,
      "size_class": "small",
      "parameter_count_basis": "Hugging Face safetensors.total for the pinned official artifact",
      "parameter_source_url": "https://huggingface.co/api/models/google/gemma-4-E2B-it/revision/3e22461f65e89153144f8adb70e3b8c2cc9845a7",
      "parameter_card_url": "https://huggingface.co/google/gemma-4-E2B-it/blob/3e22461f65e89153144f8adb70e3b8c2cc9845a7/README.md",
      "parameter_card_sha256": "3e0608a6be80e3eb040a5aa6c76707809503deb85be6360184a118fb6156526c",
      "parameter_checked_date": "2026-10-21",
      "parameter_notes": "Effective count is 2.3B; it is not an active-parameter measurement. Total includes embeddings and modality encoders."
    },
    {
      "id": "gemma4-e2b/on",
      "suite_id": "tiny",
      "model": "Gemma 4 E2B",
      "mode": "Thinking on",
      "model_key": "google/gemma-4-e2b",
      "selected_variant": "google/gemma-4-e2b@q4_k_m",
      "quantization": "Q4_K_M",
      "adapter": "native-reasoning-v1",
      "generation": {
        "context_length": 65536,
        "max_output_tokens": 32768,
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 64,
        "min_p": 0,
        "repeat_penalty": 1
      },
      "thinking": "on",
      "reasoning_effort": "not separately specified",
      "generation_seed": null,
      "generation_seed_status": "not set; native API does not expose it",
      "runtime_version": "not verified",
      "exact_gpu_placement": "not verified",
      "source_id": "tiny/gemmae2-on",
      "batch": "tiny",
      "total_parameters": 5123178051,
      "active_parameters": null,
      "effective_parameters": 2300000000,
      "size_class": "small",
      "parameter_count_basis": "Hugging Face safetensors.total for the pinned official artifact",
      "parameter_source_url": "https://huggingface.co/api/models/google/gemma-4-E2B-it/revision/3e22461f65e89153144f8adb70e3b8c2cc9845a7",
      "parameter_card_url": "https://huggingface.co/google/gemma-4-E2B-it/blob/3e22461f65e89153144f8adb70e3b8c2cc9845a7/README.md",
      "parameter_card_sha256": "3e0608a6be80e3eb040a5aa6c76707809503deb85be6360184a118fb6156526c",
      "parameter_checked_date": "2026-10-21",
      "parameter_notes": "Effective count is 2.3B; it is not an active-parameter measurement. Total includes embeddings and modality encoders."
    },
    {
      "id": "k2-horizon-3.7b/low",
      "suite_id": "tiny",
      "model": "K2 Horizon 3.7B",
      "mode": "Low reasoning effort",
      "model_key": "k2-horizon-3.7b",
      "selected_variant": "not verified",
      "quantization": "Q4_K_M",
      "adapter": "k2-creator-low-text-v1",
      "generation": {
        "context_length": 65536,
        "max_output_tokens": 32768,
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 0,
        "min_p": 0,
        "repeat_penalty": 1
      },
      "thinking": "on",
      "reasoning_effort": "not separately specified",
      "generation_seed": null,
      "generation_seed_status": "not set; native API does not expose it",
      "runtime_version": "not verified",
      "exact_gpu_placement": "not verified",
      "source_id": "tiny/k2-low",
      "batch": "tiny",
      "total_parameters": 5058255360,
      "active_parameters": 3700000000,
      "effective_parameters": null,
      "size_class": "small",
      "parameter_count_basis": "Hugging Face safetensors.total for the pinned official artifact",
      "parameter_source_url": "https://huggingface.co/api/models/IFM/K2-Horizon-3.7B/revision/9d280ff742235796925c9fe3cf243dc1beea5c08",
      "parameter_card_url": "https://huggingface.co/IFM/K2-Horizon-3.7B/blob/9d280ff742235796925c9fe3cf243dc1beea5c08/README.md",
      "parameter_card_sha256": "e0868f280e0509f75a208ca54a328808a8652d7466f5f7b76b5e8d561e9a0c3d",
      "parameter_checked_date": "2026-10-21",
      "parameter_notes": "Card reports a 3.7B core and labels 3.7B activated parameters. Artifact total also counts embeddings/output weights; active is the card convention."
    },
    {
      "id": "granite4.2-3b/off",
      "suite_id": "tiny",
      "model": "Granite 4.2 3B",
      "mode": "Thinking off",
      "model_key": "granite-4.2-3b",
      "selected_variant": "not verified",
      "quantization": "Q4_K_M",
      "adapter": "granite-creator-text-v1",
      "generation": {
        "context_length": 65536,
        "max_output_tokens": 32768,
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 0,
        "min_p": 0,
        "repeat_penalty": 1
      },
      "thinking": "off",
      "reasoning_effort": "not separately specified",
      "generation_seed": null,
      "generation_seed_status": "not set; native API does not expose it",
      "runtime_version": "not verified",
      "exact_gpu_placement": "not verified",
      "source_id": "tiny/granite-off",
      "batch": "tiny",
      "total_parameters": 3659737600,
      "active_parameters": null,
      "effective_parameters": null,
      "size_class": "small",
      "parameter_count_basis": "Hugging Face safetensors.total for the pinned official artifact",
      "parameter_source_url": "https://huggingface.co/api/models/ibm-granite/granite-4.2-3b/revision/4e9c275e232f10211ec88644d106a1f6a1b60e06",
      "parameter_card_url": "https://huggingface.co/ibm-granite/granite-4.2-3b/blob/4e9c275e232f10211ec88644d106a1f6a1b60e06/README.md",
      "parameter_card_sha256": "8f22ddce440f0c15ecb58fc9f8501b882baf5f2b7ab0d13872fe77d88c1ad792",
      "parameter_checked_date": "2026-10-21",
      "parameter_notes": "Dense model; artifact total includes separate input/output embeddings. No separate active count is published."
    },
    {
      "id": "granite4.2-3b/on",
      "suite_id": "tiny",
      "model": "Granite 4.2 3B",
      "mode": "Thinking on",
      "model_key": "granite-4.2-3b",
      "selected_variant": "not verified",
      "quantization": "Q4_K_M",
      "adapter": "granite-creator-text-v1",
      "generation": {
        "context_length": 65536,
        "max_output_tokens": 32768,
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 0,
        "min_p": 0,
        "repeat_penalty": 1
      },
      "thinking": "on",
      "reasoning_effort": "not separately specified",
      "generation_seed": null,
      "generation_seed_status": "not set; native API does not expose it",
      "runtime_version": "not verified",
      "exact_gpu_placement": "not verified",
      "source_id": "tiny/granite-on",
      "batch": "tiny",
      "total_parameters": 3659737600,
      "active_parameters": null,
      "effective_parameters": null,
      "size_class": "small",
      "parameter_count_basis": "Hugging Face safetensors.total for the pinned official artifact",
      "parameter_source_url": "https://huggingface.co/api/models/ibm-granite/granite-4.2-3b/revision/4e9c275e232f10211ec88644d106a1f6a1b60e06",
      "parameter_card_url": "https://huggingface.co/ibm-granite/granite-4.2-3b/blob/4e9c275e232f10211ec88644d106a1f6a1b60e06/README.md",
      "parameter_card_sha256": "8f22ddce440f0c15ecb58fc9f8501b882baf5f2b7ab0d13872fe77d88c1ad792",
      "parameter_checked_date": "2026-10-21",
      "parameter_notes": "Dense model; artifact total includes separate input/output embeddings. No separate active count is published."
    },
    {
      "id": "qwen3.5-4b/off",
      "suite_id": "medium",
      "model": "Qwen3.5 4B",
      "mode": "Thinking off",
      "model_key": "qwen3.5-4b",
      "selected_variant": "not verified",
      "quantization": "Q4_K_M",
      "adapter": "qwen35-official-text-v1",
      "generation": {
        "context_length": 65536,
        "max_output_tokens": 32768,
        "temperature": 0.7,
        "top_p": 0.8,
        "top_k": 20,
        "min_p": 0,
        "repeat_penalty": 1,
        "presence_penalty": 1.5
      },
      "thinking": "off",
      "reasoning_effort": "not separately specified",
      "generation_seed": null,
      "generation_seed_status": "not set; native API does not expose it",
      "runtime_version": "not verified",
      "exact_gpu_placement": "not verified",
      "source_id": "medium/qwen4-off",
      "batch": "medium",
      "total_parameters": 4659865088,
      "active_parameters": null,
      "effective_parameters": null,
      "size_class": "small",
      "parameter_count_basis": "Hugging Face safetensors.total for the pinned official artifact",
      "parameter_source_url": "https://huggingface.co/api/models/Qwen/Qwen3.5-4B/revision/851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a",
      "parameter_card_url": "https://huggingface.co/Qwen/Qwen3.5-4B/blob/851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a/README.md",
      "parameter_card_sha256": "1406be1b6b8fd8a6545870da516912804756593628a1d0fb0a7965211e82a7bb",
      "parameter_checked_date": "2026-10-21",
      "parameter_notes": "Dense language model is labelled 4B; artifact total includes the vision components. No separate active count is published."
    },
    {
      "id": "qwen3.5-4b/on",
      "suite_id": "medium",
      "model": "Qwen3.5 4B",
      "mode": "Thinking on",
      "model_key": "qwen3.5-4b",
      "selected_variant": "not verified",
      "quantization": "Q4_K_M",
      "adapter": "qwen35-official-text-v1",
      "generation": {
        "context_length": 65536,
        "max_output_tokens": 32768,
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 20,
        "min_p": 0,
        "repeat_penalty": 1,
        "presence_penalty": 1.5
      },
      "thinking": "on",
      "reasoning_effort": "not separately specified",
      "generation_seed": null,
      "generation_seed_status": "not set; native API does not expose it",
      "runtime_version": "not verified",
      "exact_gpu_placement": "not verified",
      "source_id": "medium/qwen4-on",
      "batch": "medium",
      "total_parameters": 4659865088,
      "active_parameters": null,
      "effective_parameters": null,
      "size_class": "small",
      "parameter_count_basis": "Hugging Face safetensors.total for the pinned official artifact",
      "parameter_source_url": "https://huggingface.co/api/models/Qwen/Qwen3.5-4B/revision/851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a",
      "parameter_card_url": "https://huggingface.co/Qwen/Qwen3.5-4B/blob/851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a/README.md",
      "parameter_card_sha256": "1406be1b6b8fd8a6545870da516912804756593628a1d0fb0a7965211e82a7bb",
      "parameter_checked_date": "2026-10-21",
      "parameter_notes": "Dense language model is labelled 4B; artifact total includes the vision components. No separate active count is published."
    },
    {
      "id": "granite4.2-8b/on",
      "suite_id": "medium",
      "model": "Granite 4.2 8B",
      "mode": "Thinking on",
      "model_key": "granite-4.2-8b",
      "selected_variant": "not verified",
      "quantization": "Q4_K_M",
      "adapter": "granite-creator-text-v1",
      "generation": {
        "context_length": 65536,
        "max_output_tokens": 32768,
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 0,
        "min_p": 0,
        "repeat_penalty": 1
      },
      "thinking": "on",
      "reasoning_effort": "not separately specified",
      "generation_seed": null,
      "generation_seed_status": "not set; native API does not expose it",
      "runtime_version": "not verified",
      "exact_gpu_placement": "not verified",
      "source_id": "medium/granite-on",
      "batch": "medium",
      "total_parameters": 8791592960,
      "active_parameters": null,
      "effective_parameters": null,
      "size_class": "medium",
      "parameter_count_basis": "Hugging Face safetensors.total for the pinned official artifact",
      "parameter_source_url": "https://huggingface.co/api/models/ibm-granite/granite-4.2-8b/revision/a7042144125ce5ae64f89610e8289647fc000b9a",
      "parameter_card_url": "https://huggingface.co/ibm-granite/granite-4.2-8b/blob/a7042144125ce5ae64f89610e8289647fc000b9a/README.md",
      "parameter_card_sha256": "01396b1a8596586986fea22cfaa86d0255d19113d71fe09e0b69cc2bfa4307da",
      "parameter_checked_date": "2026-10-21",
      "parameter_notes": "Dense model; artifact total includes separate input/output embeddings. No separate active count is published."
    },
    {
      "id": "granite4.2-8b/off",
      "suite_id": "medium",
      "model": "Granite 4.2 8B",
      "mode": "Thinking off",
      "model_key": "granite-4.2-8b",
      "selected_variant": "not verified",
      "quantization": "Q4_K_M",
      "adapter": "granite-creator-text-v1",
      "generation": {
        "context_length": 65536,
        "max_output_tokens": 32768,
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 0,
        "min_p": 0,
        "repeat_penalty": 1
      },
      "thinking": "off",
      "reasoning_effort": "not separately specified",
      "generation_seed": null,
      "generation_seed_status": "not set; native API does not expose it",
      "runtime_version": "not verified",
      "exact_gpu_placement": "not verified",
      "source_id": "medium/granite-off",
      "batch": "medium",
      "total_parameters": 8791592960,
      "active_parameters": null,
      "effective_parameters": null,
      "size_class": "medium",
      "parameter_count_basis": "Hugging Face safetensors.total for the pinned official artifact",
      "parameter_source_url": "https://huggingface.co/api/models/ibm-granite/granite-4.2-8b/revision/a7042144125ce5ae64f89610e8289647fc000b9a",
      "parameter_card_url": "https://huggingface.co/ibm-granite/granite-4.2-8b/blob/a7042144125ce5ae64f89610e8289647fc000b9a/README.md",
      "parameter_card_sha256": "01396b1a8596586986fea22cfaa86d0255d19113d71fe09e0b69cc2bfa4307da",
      "parameter_checked_date": "2026-10-21",
      "parameter_notes": "Dense model; artifact total includes separate input/output embeddings. No separate active count is published."
    },
    {
      "id": "lfm2.5-8b-a1b/native",
      "suite_id": "medium",
      "model": "LFM2.5-8B-A1B",
      "mode": "Native thinking",
      "model_key": "lfm2.5-8b-a1b",
      "selected_variant": "not verified",
      "quantization": "Q4_K_M",
      "adapter": "native-default-thinking-v1",
      "generation": {
        "context_length": 65536,
        "max_output_tokens": 32768,
        "temperature": 0.2,
        "top_p": 1.0,
        "top_k": 80,
        "min_p": 0,
        "repeat_penalty": 1.05
      },
      "thinking": "on",
      "reasoning_effort": "not separately specified",
      "generation_seed": null,
      "generation_seed_status": "not set; native API does not expose it",
      "runtime_version": "not verified",
      "exact_gpu_placement": "not verified",
      "source_id": "medium/lfm8-on",
      "batch": "medium",
      "total_parameters": 8467856832,
      "active_parameters": 1500000000,
      "effective_parameters": null,
      "size_class": "medium",
      "parameter_count_basis": "Hugging Face safetensors.total for the pinned official artifact",
      "parameter_source_url": "https://huggingface.co/api/models/LiquidAI/LFM2.5-8B-A1B/revision/5dd22602c2e9f6a097b1de4c4efe0658b605015c",
      "parameter_card_url": "https://huggingface.co/LiquidAI/LFM2.5-8B-A1B/blob/5dd22602c2e9f6a097b1de4c4efe0658b605015c/README.md",
      "parameter_card_sha256": "eb40f5396ef5200a14240f8c0523b269bcc114f40b153ac2fa9017dad391e59b",
      "parameter_checked_date": "2026-10-21",
      "parameter_notes": "Card reports 8.3B total and 1.5B active; active count is not the A1B label rounded to an integer."
    },
    {
      "id": "qwen3.5-9b/off",
      "suite_id": "medium",
      "model": "Qwen3.5 9B",
      "mode": "Thinking off",
      "model_key": "qwen/qwen3.5-9b",
      "selected_variant": "qwen/qwen3.5-9b@q4_k_m",
      "quantization": "Q4_K_M",
      "adapter": "qwen35-official-text-v1",
      "generation": {
        "context_length": 65536,
        "max_output_tokens": 32768,
        "temperature": 0.7,
        "top_p": 0.8,
        "top_k": 20,
        "min_p": 0,
        "repeat_penalty": 1,
        "presence_penalty": 1.5
      },
      "thinking": "off",
      "reasoning_effort": "not separately specified",
      "generation_seed": null,
      "generation_seed_status": "not set; native API does not expose it",
      "runtime_version": "not verified",
      "exact_gpu_placement": "not verified",
      "source_id": "medium/qwen9-off",
      "batch": "medium",
      "total_parameters": 9653104368,
      "active_parameters": null,
      "effective_parameters": null,
      "size_class": "medium",
      "parameter_count_basis": "Hugging Face safetensors.total for the pinned official artifact",
      "parameter_source_url": "https://huggingface.co/api/models/Qwen/Qwen3.5-9B/revision/c202236235762e1c871ad0ccb60c8ee5ba337b9a",
      "parameter_card_url": "https://huggingface.co/Qwen/Qwen3.5-9B/blob/c202236235762e1c871ad0ccb60c8ee5ba337b9a/README.md",
      "parameter_card_sha256": "c5f5a8c2dddab69cfbf05279235aa5fddb137939a06539c4c7637aa900fef6d0",
      "parameter_checked_date": "2026-10-21",
      "parameter_notes": "Dense language model is labelled 9B; artifact total includes the vision components. No separate active count is published."
    },
    {
      "id": "qwen3.5-9b/on",
      "suite_id": "medium",
      "model": "Qwen3.5 9B",
      "mode": "Thinking on",
      "model_key": "qwen/qwen3.5-9b",
      "selected_variant": "qwen/qwen3.5-9b@q4_k_m",
      "quantization": "Q4_K_M",
      "adapter": "qwen35-official-text-v1",
      "generation": {
        "context_length": 65536,
        "max_output_tokens": 32768,
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 20,
        "min_p": 0,
        "repeat_penalty": 1,
        "presence_penalty": 1.5
      },
      "thinking": "on",
      "reasoning_effort": "not separately specified",
      "generation_seed": null,
      "generation_seed_status": "not set; native API does not expose it",
      "runtime_version": "not verified",
      "exact_gpu_placement": "not verified",
      "source_id": "medium/qwen9-on",
      "batch": "medium",
      "total_parameters": 9653104368,
      "active_parameters": null,
      "effective_parameters": null,
      "size_class": "medium",
      "parameter_count_basis": "Hugging Face safetensors.total for the pinned official artifact",
      "parameter_source_url": "https://huggingface.co/api/models/Qwen/Qwen3.5-9B/revision/c202236235762e1c871ad0ccb60c8ee5ba337b9a",
      "parameter_card_url": "https://huggingface.co/Qwen/Qwen3.5-9B/blob/c202236235762e1c871ad0ccb60c8ee5ba337b9a/README.md",
      "parameter_card_sha256": "c5f5a8c2dddab69cfbf05279235aa5fddb137939a06539c4c7637aa900fef6d0",
      "parameter_checked_date": "2026-10-21",
      "parameter_notes": "Dense language model is labelled 9B; artifact total includes the vision components. No separate active count is published."
    },
    {
      "id": "granite4.2-30b/on",
      "suite_id": "large",
      "model": "Granite 4.2 30B",
      "mode": "Thinking on",
      "model_key": "granite-4.2-30b",
      "selected_variant": "not verified",
      "quantization": "Q4_K_M",
      "adapter": "granite-creator-text-v1",
      "generation": {
        "context_length": 65536,
        "max_output_tokens": 32768,
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 0,
        "min_p": 0,
        "repeat_penalty": 1.0
      },
      "thinking": "on",
      "reasoning_effort": "not separately specified",
      "generation_seed": null,
      "generation_seed_status": "not set; native API does not expose it",
      "runtime_version": "not verified",
      "exact_gpu_placement": "not verified",
      "source_id": "large/granite-on",
      "batch": "large",
      "total_parameters": 29276770304,
      "active_parameters": null,
      "effective_parameters": null,
      "size_class": "large",
      "parameter_count_basis": "Hugging Face safetensors.total for the pinned official artifact",
      "parameter_source_url": "https://huggingface.co/api/models/ibm-granite/granite-4.2-30b/revision/9e668ce1c538387ef24d3644e9b0606647762636",
      "parameter_card_url": "https://huggingface.co/ibm-granite/granite-4.2-30b/blob/9e668ce1c538387ef24d3644e9b0606647762636/README.md",
      "parameter_card_sha256": "10d0e2b083c888814dc92a6308ca34d1c6fe01f344b54ff54905a8cbe5a80b06",
      "parameter_checked_date": "2026-10-21",
      "parameter_notes": "Dense model; artifact total includes separate input/output embeddings. No separate active count is published."
    },
    {
      "id": "granite4.2-30b/off",
      "suite_id": "large",
      "model": "Granite 4.2 30B",
      "mode": "Thinking off",
      "model_key": "granite-4.2-30b",
      "selected_variant": "not verified",
      "quantization": "Q4_K_M",
      "adapter": "granite-creator-text-v1",
      "generation": {
        "context_length": 65536,
        "max_output_tokens": 32768,
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 0,
        "min_p": 0,
        "repeat_penalty": 1.0
      },
      "thinking": "off",
      "reasoning_effort": "not separately specified",
      "generation_seed": null,
      "generation_seed_status": "not set; native API does not expose it",
      "runtime_version": "not verified",
      "exact_gpu_placement": "not verified",
      "source_id": "large/granite-off",
      "batch": "large",
      "total_parameters": 29276770304,
      "active_parameters": null,
      "effective_parameters": null,
      "size_class": "large",
      "parameter_count_basis": "Hugging Face safetensors.total for the pinned official artifact",
      "parameter_source_url": "https://huggingface.co/api/models/ibm-granite/granite-4.2-30b/revision/9e668ce1c538387ef24d3644e9b0606647762636",
      "parameter_card_url": "https://huggingface.co/ibm-granite/granite-4.2-30b/blob/9e668ce1c538387ef24d3644e9b0606647762636/README.md",
      "parameter_card_sha256": "10d0e2b083c888814dc92a6308ca34d1c6fe01f344b54ff54905a8cbe5a80b06",
      "parameter_checked_date": "2026-10-21",
      "parameter_notes": "Dense model; artifact total includes separate input/output embeddings. No separate active count is published."
    },
    {
      "id": "lfm2.5-24b-a2b/native",
      "suite_id": "large",
      "model": "LFM2.5-24B-A2B",
      "mode": "Native thinking",
      "model_key": "lfm2.5-24b-a2b",
      "selected_variant": "not verified",
      "quantization": "Q4_K_M",
      "adapter": "native-default-thinking-v1",
      "generation": {
        "context_length": 65536,
        "max_output_tokens": 32768,
        "temperature": 0.1,
        "top_p": 1.0,
        "top_k": 50,
        "min_p": 0,
        "repeat_penalty": 1.05
      },
      "thinking": "on",
      "reasoning_effort": "not separately specified",
      "generation_seed": null,
      "generation_seed_status": "not set; native API does not expose it",
      "runtime_version": "not verified",
      "exact_gpu_placement": "not verified",
      "source_id": "large/lfm24-on",
      "batch": "large",
      "total_parameters": 23971585280,
      "active_parameters": 2450000000,
      "effective_parameters": null,
      "size_class": "large",
      "parameter_count_basis": "Hugging Face safetensors.total for the pinned official artifact",
      "parameter_source_url": "https://huggingface.co/api/models/LiquidAI/LFM2.5-24B-A2B/revision/3fcfdbfa1a1b6d8166846a6323e69f1e1254440d",
      "parameter_card_url": "https://huggingface.co/LiquidAI/LFM2.5-24B-A2B/blob/3fcfdbfa1a1b6d8166846a6323e69f1e1254440d/README.md",
      "parameter_card_sha256": "e5e174cc7f03caf4621e2372395ea558765bf5cddbcf052a6e591090b491af69",
      "parameter_checked_date": "2026-10-21",
      "parameter_notes": "Card reports 24.0B total and 2.45B active."
    },
    {
      "id": "qwen3.8-27b/off",
      "suite_id": "large",
      "model": "Qwen3.8 27B",
      "mode": "Thinking off",
      "model_key": "qwen/qwen3.8-27b",
      "selected_variant": "qwen/qwen3.8-27b@q4_k_m",
      "quantization": "Q4_K_M",
      "adapter": "qwen38-creator-text-v1",
      "generation": {
        "context_length": 65536,
        "max_output_tokens": 32768,
        "temperature": 0.7,
        "top_p": 0.8,
        "top_k": 20,
        "min_p": 0,
        "repeat_penalty": 1.0,
        "presence_penalty": 1.5
      },
      "thinking": "off",
      "reasoning_effort": "off",
      "generation_seed": null,
      "generation_seed_status": "not set; native API does not expose it",
      "runtime_version": "not verified",
      "exact_gpu_placement": "not verified",
      "source_id": "large/qwen27-off",
      "batch": "large",
      "total_parameters": 27781427952,
      "active_parameters": null,
      "effective_parameters": null,
      "size_class": "large",
      "parameter_count_basis": "Hugging Face safetensors.total for the pinned official artifact",
      "parameter_source_url": "https://huggingface.co/api/models/Qwen/Qwen3.8-27B/revision/1d4bf0f2ff6012fd82039f2fa52739d0dd7c60c0",
      "parameter_card_url": "https://huggingface.co/Qwen/Qwen3.8-27B/blob/1d4bf0f2ff6012fd82039f2fa52739d0dd7c60c0/README.md",
      "parameter_card_sha256": "57e4bdb258ee1a7d2635c5174ebd4e56abe392505cdb5f8bbb356b0dc4293641",
      "parameter_checked_date": "2026-10-21",
      "parameter_notes": "Dense language model is labelled 27B; artifact total includes the vision components. No separate active count is published."
    },
    {
      "id": "qwen3.8-27b/on",
      "suite_id": "large",
      "model": "Qwen3.8 27B",
      "mode": "Low thinking effort",
      "model_key": "qwen/qwen3.8-27b",
      "selected_variant": "qwen/qwen3.8-27b@q4_k_m",
      "quantization": "Q4_K_M",
      "adapter": "qwen38-creator-text-v1",
      "generation": {
        "context_length": 65536,
        "max_output_tokens": 32768,
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 20,
        "min_p": 0,
        "repeat_penalty": 1.0,
        "presence_penalty": 0.0
      },
      "thinking": "on",
      "reasoning_effort": "low",
      "generation_seed": null,
      "generation_seed_status": "not set; native API does not expose it",
      "runtime_version": "not verified",
      "exact_gpu_placement": "not verified",
      "source_id": "large/qwen27-on",
      "batch": "large",
      "total_parameters": 27781427952,
      "active_parameters": null,
      "effective_parameters": null,
      "size_class": "large",
      "parameter_count_basis": "Hugging Face safetensors.total for the pinned official artifact",
      "parameter_source_url": "https://huggingface.co/api/models/Qwen/Qwen3.8-27B/revision/1d4bf0f2ff6012fd82039f2fa52739d0dd7c60c0",
      "parameter_card_url": "https://huggingface.co/Qwen/Qwen3.8-27B/blob/1d4bf0f2ff6012fd82039f2fa52739d0dd7c60c0/README.md",
      "parameter_card_sha256": "57e4bdb258ee1a7d2635c5174ebd4e56abe392505cdb5f8bbb356b0dc4293641",
      "parameter_checked_date": "2026-10-21",
      "parameter_notes": "Dense language model is labelled 27B; artifact total includes the vision components. No separate active count is published."
    },
    {
      "id": "gemma4-26b-a4b/off",
      "suite_id": "large",
      "model": "Gemma 4 26B-A4B",
      "mode": "Thinking off",
      "model_key": "google/gemma-4-26b-a4b",
      "selected_variant": "google/gemma-4-26b-a4b@q4_k_m",
      "quantization": "Q4_K_M",
      "adapter": "native-reasoning-v1",
      "generation": {
        "context_length": 65536,
        "max_output_tokens": 32768,
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 64,
        "min_p": 0,
        "repeat_penalty": 1.0
      },
      "thinking": "off",
      "reasoning_effort": "not separately specified",
      "generation_seed": null,
      "generation_seed_status": "not set; native API does not expose it",
      "runtime_version": "not verified",
      "exact_gpu_placement": "not verified",
      "source_id": "large/gemma26-off",
      "batch": "large",
      "total_parameters": 25805936206,
      "active_parameters": 3800000000,
      "effective_parameters": null,
      "size_class": "large",
      "parameter_count_basis": "Hugging Face safetensors.total for the pinned official artifact",
      "parameter_source_url": "https://huggingface.co/api/models/google/gemma-4-26B-A4B-it/revision/4d7ae4984b7db7de8f8457170b3f1a419ee76d52",
      "parameter_card_url": "https://huggingface.co/google/gemma-4-26B-A4B-it/blob/4d7ae4984b7db7de8f8457170b3f1a419ee76d52/README.md",
      "parameter_card_sha256": "4f463225c4fe49ec120115b0f9bb6713ff1556293a909903ffdae3df7bf974da",
      "parameter_checked_date": "2026-10-21",
      "parameter_notes": "Card gives 25.2B language-model total and 3.8B active; artifact total also includes the vision encoder."
    },
    {
      "id": "gemma4-26b-a4b/on",
      "suite_id": "large",
      "model": "Gemma 4 26B-A4B",
      "mode": "Thinking on",
      "model_key": "google/gemma-4-26b-a4b",
      "selected_variant": "google/gemma-4-26b-a4b@q4_k_m",
      "quantization": "Q4_K_M",
      "adapter": "native-reasoning-v1",
      "generation": {
        "context_length": 65536,
        "max_output_tokens": 32768,
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 64,
        "min_p": 0,
        "repeat_penalty": 1.0
      },
      "thinking": "on",
      "reasoning_effort": "not separately specified",
      "generation_seed": null,
      "generation_seed_status": "not set; native API does not expose it",
      "runtime_version": "not verified",
      "exact_gpu_placement": "not verified",
      "source_id": "large/gemma26-on",
      "batch": "large",
      "total_parameters": 25805936206,
      "active_parameters": 3800000000,
      "effective_parameters": null,
      "size_class": "large",
      "parameter_count_basis": "Hugging Face safetensors.total for the pinned official artifact",
      "parameter_source_url": "https://huggingface.co/api/models/google/gemma-4-26B-A4B-it/revision/4d7ae4984b7db7de8f8457170b3f1a419ee76d52",
      "parameter_card_url": "https://huggingface.co/google/gemma-4-26B-A4B-it/blob/4d7ae4984b7db7de8f8457170b3f1a419ee76d52/README.md",
      "parameter_card_sha256": "4f463225c4fe49ec120115b0f9bb6713ff1556293a909903ffdae3df7bf974da",
      "parameter_checked_date": "2026-10-21",
      "parameter_notes": "Card gives 25.2B language-model total and 3.8B active; artifact total also includes the vision encoder."
    },
    {
      "id": "glm4.7-flash/off",
      "suite_id": "large",
      "model": "GLM-4.7 Flash",
      "mode": "Thinking off",
      "model_key": "zai-org/glm-4.7-flash",
      "selected_variant": "zai-org/glm-4.7-flash@q4_k_m",
      "quantization": "Q4_K_M",
      "adapter": "glm47-creator-text-v1",
      "generation": {
        "context_length": 65536,
        "max_output_tokens": 32768,
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 0,
        "min_p": 0,
        "repeat_penalty": 1.0,
        "presence_penalty": 0.0
      },
      "thinking": "off",
      "reasoning_effort": "not separately specified",
      "generation_seed": null,
      "generation_seed_status": "not set; native API does not expose it",
      "runtime_version": "not verified",
      "exact_gpu_placement": "not verified",
      "source_id": "large/glm30-off",
      "batch": "large",
      "total_parameters": 31221488576,
      "active_parameters": 3000000000,
      "effective_parameters": null,
      "size_class": "large",
      "parameter_count_basis": "Hugging Face safetensors.total for the pinned official artifact",
      "parameter_source_url": "https://huggingface.co/api/models/zai-org/GLM-4.7-Flash/revision/7dd20894a642a0aa287e9827cb1a1f7f91386b67",
      "parameter_card_url": "https://huggingface.co/zai-org/GLM-4.7-Flash/blob/7dd20894a642a0aa287e9827cb1a1f7f91386b67/README.md",
      "parameter_card_sha256": "7bfcbced42741ceb3fff8809f8d6c84ada9696850e257c4e53e6c2fe9ef4f9bf",
      "parameter_checked_date": "2026-10-21",
      "parameter_notes": "Card labels the model 30B-A3B; artifact total is about 31.22B."
    },
    {
      "id": "glm4.7-flash/on",
      "suite_id": "large",
      "model": "GLM-4.7 Flash",
      "mode": "Thinking on",
      "model_key": "zai-org/glm-4.7-flash",
      "selected_variant": "zai-org/glm-4.7-flash@q4_k_m",
      "quantization": "Q4_K_M",
      "adapter": "glm47-creator-text-v1",
      "generation": {
        "context_length": 65536,
        "max_output_tokens": 32768,
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 0,
        "min_p": 0,
        "repeat_penalty": 1.0,
        "presence_penalty": 0.0
      },
      "thinking": "on",
      "reasoning_effort": "not separately specified",
      "generation_seed": null,
      "generation_seed_status": "not set; native API does not expose it",
      "runtime_version": "not verified",
      "exact_gpu_placement": "not verified",
      "source_id": "large/glm30-on",
      "batch": "large",
      "total_parameters": 31221488576,
      "active_parameters": 3000000000,
      "effective_parameters": null,
      "size_class": "large",
      "parameter_count_basis": "Hugging Face safetensors.total for the pinned official artifact",
      "parameter_source_url": "https://huggingface.co/api/models/zai-org/GLM-4.7-Flash/revision/7dd20894a642a0aa287e9827cb1a1f7f91386b67",
      "parameter_card_url": "https://huggingface.co/zai-org/GLM-4.7-Flash/blob/7dd20894a642a0aa287e9827cb1a1f7f91386b67/README.md",
      "parameter_card_sha256": "7bfcbced42741ceb3fff8809f8d6c84ada9696850e257c4e53e6c2fe9ef4f9bf",
      "parameter_checked_date": "2026-10-21",
      "parameter_notes": "Card labels the model 30B-A3B; artifact total is about 31.22B."
    },
    {
      "id": "muse-glimmer/low",
      "suite_id": "large",
      "model": "Muse Glimmer",
      "mode": "Low thinking effort",
      "model_key": "meta/muse-glimmer",
      "selected_variant": "meta/muse-glimmer@q4_k_m",
      "quantization": "Q4_K_M",
      "adapter": "native-effort-v1",
      "generation": {
        "context_length": 65536,
        "max_output_tokens": 32768,
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 64,
        "min_p": 0,
        "repeat_penalty": 1.0
      },
      "thinking": "on",
      "reasoning_effort": "low",
      "generation_seed": null,
      "generation_seed_status": "not set; native API does not expose it",
      "runtime_version": "not verified",
      "exact_gpu_placement": "not verified",
      "source_id": "large/muse-low",
      "batch": "large",
      "total_parameters": 29776626688,
      "active_parameters": null,
      "effective_parameters": null,
      "size_class": "large",
      "parameter_count_basis": "Hugging Face safetensors.total for the pinned official artifact",
      "parameter_source_url": "https://huggingface.co/api/models/meta-models/Muse-Glimmer-30B/revision/7af58019cf472021aae5480731c9ae8e141b56c7",
      "parameter_card_url": "https://huggingface.co/meta-models/Muse-Glimmer-30B/blob/7af58019cf472021aae5480731c9ae8e141b56c7/README.md",
      "parameter_card_sha256": "2ba646dcbe5b3b97daedc59e64ef19505312fa99b6598cacc0b82d93f4d74163",
      "parameter_checked_date": "2026-10-21",
      "parameter_notes": "Card describes a dense model and rounds total including vision to about 29.6B; artifact metadata gives about 29.78B. No separate active count is published."
    }
  ],
  "cells": [
    {
      "configuration": "gemma4-e4b/off",
      "condition": "no_source",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 35.9,
        "fully_correct": 7,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.3015277777777778,
        "answers_with_material_errors": 111,
        "material_error_assessed": 120,
        "pending_reviews": 0,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 517.6083333333333,
        "mean_client_seconds": 5.604219952783144
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 5.604219952783144,
        "median_seconds": 5.011316405492835,
        "p90_seconds": 8.151041434000945,
        "p95_seconds": 10.610105722997105,
        "p99_seconds": 15.031943841000611,
        "max_seconds": 21.645217498997226,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 5.011316405492835,
        "final_returned_p95_seconds": 10.610105722997105,
        "first_token_timing_cases": 120,
        "first_token_p95_seconds": 0.204363,
        "observed_above_seconds": {
          "120": {
            "count": 0,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 3.9278582879924215,
        "p75_seconds": 6.702583901002072,
        "final_answer_completion": {
          "count": 120,
          "mean": 5.604219952783144,
          "median": 5.011316405492835,
          "p95": 10.610105722997105,
          "maximum": 21.645217498997226,
          "p25": 3.9278582879924215,
          "p75": 6.702583901002072
        },
        "above_30_seconds": 0,
        "bins": [
          58,
          54,
          7,
          1,
          0
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 517.6083333333333,
        "median": 466.0,
        "p95": 911,
        "maximum": 1448
      },
      "input_tokens": {
        "count": 120,
        "mean": 122.33333333333333,
        "median": 114.0,
        "p95": 181,
        "maximum": 292
      },
      "thinking_tokens": {
        "count": 120,
        "mean": 0,
        "median": 0.0,
        "p95": 0,
        "maximum": 0
      },
      "server_tokens_per_second": {
        "count": 120,
        "mean": 133.90520917858413,
        "median": 134.47019797048185,
        "p95": 135.44155696816824,
        "maximum": 135.9396015794887
      },
      "first_token_seconds": {
        "count": 120,
        "mean": 0.08948276666666667,
        "median": 0.0572235,
        "p95": 0.204363,
        "maximum": 0.336395
      },
      "no_final": 0
    },
    {
      "configuration": "gemma4-e4b/off",
      "condition": "supplied_passage",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 87.11666666666666,
        "fully_correct": 78,
        "profile_assessed": 119,
        "mean_fully_met_claim_coverage": 0.8484722222222222,
        "answers_with_material_errors": 36,
        "material_error_assessed": 119,
        "pending_reviews": 1,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 383.68333333333334,
        "mean_client_seconds": 4.342100253925612
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 4.342100253925612,
        "median_seconds": 3.9901296970056137,
        "p90_seconds": 7.080524860008154,
        "p95_seconds": 8.767960596000194,
        "p99_seconds": 10.123264835994632,
        "max_seconds": 11.100527478993172,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 3.9901296970056137,
        "final_returned_p95_seconds": 8.767960596000194,
        "first_token_timing_cases": 120,
        "first_token_p95_seconds": 0.31690199999999996,
        "observed_above_seconds": {
          "120": {
            "count": 0,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 2.95872265601065,
        "p75_seconds": 5.111951914004749,
        "final_answer_completion": {
          "count": 120,
          "mean": 4.342100253925612,
          "median": 3.9901296970056137,
          "p95": 8.767960596000194,
          "maximum": 11.100527478993172,
          "p25": 2.95872265601065,
          "p75": 5.111951914004749
        },
        "above_30_seconds": 0,
        "bins": [
          85,
          32,
          3,
          0,
          0
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 383.68333333333334,
        "median": 331.0,
        "p95": 892,
        "maximum": 1319
      },
      "input_tokens": {
        "count": 120,
        "mean": 357.1666666666667,
        "median": 343.5,
        "p95": 575,
        "maximum": 999
      },
      "thinking_tokens": {
        "count": 120,
        "mean": 0,
        "median": 0.0,
        "p95": 0,
        "maximum": 0
      },
      "server_tokens_per_second": {
        "count": 120,
        "mean": 133.32863204484724,
        "median": 133.75490822189965,
        "p95": 134.76205728125854,
        "maximum": 135.30124980596332
      },
      "first_token_seconds": {
        "count": 120,
        "mean": 0.14091244166666667,
        "median": 0.100392,
        "p95": 0.31690199999999996,
        "maximum": 0.7095629999999999
      },
      "no_final": 0
    },
    {
      "configuration": "gemma4-e4b/on",
      "condition": "no_source",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 39.266666666666666,
        "fully_correct": 11,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.33791666666666664,
        "answers_with_material_errors": 105,
        "material_error_assessed": 120,
        "pending_reviews": 2,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 1058.3166666666666,
        "mean_client_seconds": 9.414645599692154
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 9.414645599692154,
        "median_seconds": 9.312436531996354,
        "p90_seconds": 11.989284109993605,
        "p95_seconds": 13.034773176012095,
        "p99_seconds": 15.148238533001859,
        "max_seconds": 15.985178886010544,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 9.312436531996354,
        "final_returned_p95_seconds": 13.034773176012095,
        "first_token_timing_cases": 120,
        "first_token_p95_seconds": 0.17994200000000002,
        "observed_above_seconds": {
          "120": {
            "count": 0,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 7.977075734001119,
        "p75_seconds": 10.960555558995111,
        "final_answer_completion": {
          "count": 120,
          "mean": 9.414645599692154,
          "median": 9.312436531996354,
          "p95": 13.034773176012095,
          "maximum": 15.985178886010544,
          "p25": 7.977075734001119,
          "p75": 10.960555558995111
        },
        "above_30_seconds": 0,
        "bins": [
          2,
          66,
          52,
          0,
          0
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 1058.3166666666666,
        "median": 1037.0,
        "p95": 1439,
        "maximum": 1812
      },
      "input_tokens": {
        "count": 120,
        "mean": 124.33333333333333,
        "median": 116.0,
        "p95": 183,
        "maximum": 294
      },
      "thinking_tokens": {
        "count": 120,
        "mean": 734.8083333333333,
        "median": 736.0,
        "p95": 1056,
        "maximum": 1326
      },
      "server_tokens_per_second": {
        "count": 120,
        "mean": 133.11994704429705,
        "median": 133.31215827286997,
        "p95": 134.58588224102053,
        "maximum": 135.39883958759444
      },
      "first_token_seconds": {
        "count": 120,
        "mean": 0.085348875,
        "median": 0.0597655,
        "p95": 0.17994200000000002,
        "maximum": 0.361241
      },
      "no_final": 0
    },
    {
      "configuration": "gemma4-e4b/on",
      "condition": "supplied_passage",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 93.65833333333333,
        "fully_correct": 91,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.91375,
        "answers_with_material_errors": 23,
        "material_error_assessed": 120,
        "pending_reviews": 0,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 801.6916666666667,
        "mean_client_seconds": 7.350844491072947
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 7.350844491072947,
        "median_seconds": 7.0208621914935065,
        "p90_seconds": 9.934603402012726,
        "p95_seconds": 11.056239970988827,
        "p99_seconds": 13.79085455800032,
        "max_seconds": 16.293478806997882,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 7.0208621914935065,
        "final_returned_p95_seconds": 11.056239970988827,
        "first_token_timing_cases": 120,
        "first_token_p95_seconds": 0.306333,
        "observed_above_seconds": {
          "120": {
            "count": 0,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 5.995619124994846,
        "p75_seconds": 8.448429267999018,
        "final_answer_completion": {
          "count": 120,
          "mean": 7.350844491072947,
          "median": 7.0208621914935065,
          "p95": 11.056239970988827,
          "maximum": 16.293478806997882,
          "p25": 5.995619124994846,
          "p75": 8.448429267999018
        },
        "above_30_seconds": 0,
        "bins": [
          13,
          95,
          12,
          0,
          0
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 801.6916666666667,
        "median": 776.0,
        "p95": 1279,
        "maximum": 1881
      },
      "input_tokens": {
        "count": 120,
        "mean": 359.1666666666667,
        "median": 345.5,
        "p95": 577,
        "maximum": 1001
      },
      "thinking_tokens": {
        "count": 120,
        "mean": 532.9083333333333,
        "median": 533.5,
        "p95": 876,
        "maximum": 1341
      },
      "server_tokens_per_second": {
        "count": 120,
        "mean": 133.28725227149147,
        "median": 133.3475955788746,
        "p95": 134.6043072675431,
        "maximum": 136.01653680881853
      },
      "first_token_seconds": {
        "count": 120,
        "mean": 0.13695515,
        "median": 0.102504,
        "p95": 0.306333,
        "maximum": 0.379988
      },
      "no_final": 0
    },
    {
      "configuration": "lfm2.5-2.6b/native",
      "condition": "no_source",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 28.4,
        "fully_correct": 1,
        "profile_assessed": 119,
        "mean_fully_met_claim_coverage": 0.23180555555555554,
        "answers_with_material_errors": 118,
        "material_error_assessed": 119,
        "pending_reviews": 8,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 2392,
        "mean_client_seconds": 9.818030036818284
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 9.818030036818284,
        "median_seconds": 8.30514467699686,
        "p90_seconds": 18.262600780988578,
        "p95_seconds": 23.245326806005323,
        "p99_seconds": 27.825941012997646,
        "max_seconds": 30.81625905999681,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 8.30514467699686,
        "final_returned_p95_seconds": 23.245326806005323,
        "first_token_timing_cases": 120,
        "first_token_p95_seconds": 0.037697,
        "observed_above_seconds": {
          "120": {
            "count": 0,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 5.7175636330212,
        "p75_seconds": 11.722402907005744,
        "final_answer_completion": {
          "count": 120,
          "mean": 9.818030036818284,
          "median": 8.30514467699686,
          "p95": 23.245326806005323,
          "maximum": 30.81625905999681,
          "p25": 5.7175636330212,
          "p75": 11.722402907005744
        },
        "above_30_seconds": 1,
        "bins": [
          16,
          60,
          36,
          7,
          1
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 2392,
        "median": 2016.5,
        "p95": 5608,
        "maximum": 7506
      },
      "input_tokens": {
        "count": 120,
        "mean": 125.6,
        "median": 117.0,
        "p95": 182,
        "maximum": 302
      },
      "thinking_tokens": {
        "count": 120,
        "mean": 1948.9083333333333,
        "median": 1587.5,
        "p95": 5120,
        "maximum": 7051
      },
      "server_tokens_per_second": {
        "count": 120,
        "mean": 247.66187037053248,
        "median": 247.8866057309287,
        "p95": 253.51327614889644,
        "maximum": 253.88756069453208
      },
      "first_token_seconds": {
        "count": 120,
        "mean": 0.027623858333333334,
        "median": 0.025594,
        "p95": 0.037697,
        "maximum": 0.043299
      },
      "no_final": 0
    },
    {
      "configuration": "lfm2.5-2.6b/native",
      "condition": "supplied_passage",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 74.03333333333333,
        "fully_correct": 32,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.7061111111111111,
        "answers_with_material_errors": 87,
        "material_error_assessed": 120,
        "pending_reviews": 1,
        "operationally_flagged_answers": 2,
        "output_ceiling_hits": 2,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 1733.9666666666667,
        "mean_client_seconds": 7.186745574226127
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 7.186745574226127,
        "median_seconds": 4.9920281510130735,
        "p90_seconds": 14.25175850300002,
        "p95_seconds": 21.39023340199492,
        "p99_seconds": 33.112668914021924,
        "max_seconds": 34.40301293801167,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 2,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 4.9920281510130735,
        "final_returned_p95_seconds": 21.39023340199492,
        "first_token_timing_cases": 120,
        "first_token_p95_seconds": 0.075354,
        "observed_above_seconds": {
          "120": {
            "count": 0,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 3.5915746570099145,
        "p75_seconds": 7.964336066012038,
        "final_answer_completion": {
          "count": 120,
          "mean": 7.186745574226127,
          "median": 4.9920281510130735,
          "p95": 21.39023340199492,
          "maximum": 34.40301293801167,
          "p25": 3.5915746570099145,
          "p75": 7.964336066012038
        },
        "above_30_seconds": 3,
        "bins": [
          60,
          39,
          14,
          4,
          3
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 1733.9666666666667,
        "median": 1211.0,
        "p95": 5153,
        "maximum": 8192
      },
      "input_tokens": {
        "count": 120,
        "mean": 343.3833333333333,
        "median": 333.0,
        "p95": 547,
        "maximum": 962
      },
      "thinking_tokens": {
        "count": 120,
        "mean": 1337.2916666666667,
        "median": 782.5,
        "p95": 4427,
        "maximum": 7968
      },
      "server_tokens_per_second": {
        "count": 120,
        "mean": 247.95628686900147,
        "median": 249.19722947242758,
        "p95": 253.64509597520498,
        "maximum": 254.28273300620634
      },
      "first_token_seconds": {
        "count": 120,
        "mean": 0.04637028333333333,
        "median": 0.042829000000000006,
        "p95": 0.075354,
        "maximum": 0.106171
      },
      "no_final": 0
    },
    {
      "configuration": "gemma4-e2b/off",
      "condition": "no_source",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 29.566666666666666,
        "fully_correct": 8,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.24180555555555555,
        "answers_with_material_errors": 109,
        "material_error_assessed": 120,
        "pending_reviews": 1,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 464.825,
        "mean_client_seconds": 2.3145293871925485
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 2.3145293871925485,
        "median_seconds": 2.089287222508574,
        "p90_seconds": 3.4263049159781076,
        "p95_seconds": 3.953539179987274,
        "p99_seconds": 4.970880577981006,
        "max_seconds": 8.249619626032654,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 2.089287222508574,
        "final_returned_p95_seconds": 3.953539179987274,
        "first_token_timing_cases": 120,
        "first_token_p95_seconds": 0.054,
        "observed_above_seconds": {
          "120": {
            "count": 0,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 1.6495677439961582,
        "p75_seconds": 2.7630482550011948,
        "final_answer_completion": {
          "count": 120,
          "mean": 2.3145293871925485,
          "median": 2.089287222508574,
          "p95": 3.953539179987274,
          "maximum": 8.249619626032654,
          "p25": 1.6495677439961582,
          "p75": 2.7630482550011948
        },
        "above_30_seconds": 0,
        "bins": [
          119,
          1,
          0,
          0,
          0
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 464.825,
        "median": 421.5,
        "p95": 772,
        "maximum": 1690
      },
      "input_tokens": {
        "count": 120,
        "mean": 122.33333333333333,
        "median": 114.0,
        "p95": 181,
        "maximum": 292
      },
      "thinking_tokens": {
        "count": 120,
        "mean": 0,
        "median": 0.0,
        "p95": 0,
        "maximum": 0
      },
      "server_tokens_per_second": {
        "count": 120,
        "mean": 207.84264211372872,
        "median": 207.94488563550817,
        "p95": 210.38072196397118,
        "maximum": 211.3734196534171
      },
      "first_token_seconds": {
        "count": 120,
        "mean": 0.04626144166666667,
        "median": 0.0405905,
        "p95": 0.054,
        "maximum": 0.493073
      },
      "no_final": 0
    },
    {
      "configuration": "gemma4-e2b/off",
      "condition": "supplied_passage",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 79.84166666666667,
        "fully_correct": 58,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.7609722222222223,
        "answers_with_material_errors": 57,
        "material_error_assessed": 120,
        "pending_reviews": 0,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 380.0416666666667,
        "mean_client_seconds": 1.9184900485633989
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 1.9184900485633989,
        "median_seconds": 1.7609122325084172,
        "p90_seconds": 3.2197031659889035,
        "p95_seconds": 4.364872126025148,
        "p99_seconds": 5.043119853013195,
        "max_seconds": 5.117597347998526,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 1.7609122325084172,
        "final_returned_p95_seconds": 4.364872126025148,
        "first_token_timing_cases": 120,
        "first_token_p95_seconds": 0.104512,
        "observed_above_seconds": {
          "120": {
            "count": 0,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 1.1391726629808545,
        "p75_seconds": 2.367357023002114,
        "final_answer_completion": {
          "count": 120,
          "mean": 1.9184900485633989,
          "median": 1.7609122325084172,
          "p95": 4.364872126025148,
          "maximum": 5.117597347998526,
          "p25": 1.1391726629808545,
          "p75": 2.367357023002114
        },
        "above_30_seconds": 0,
        "bins": [
          118,
          2,
          0,
          0,
          0
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 380.0416666666667,
        "median": 345.0,
        "p95": 875,
        "maximum": 1041
      },
      "input_tokens": {
        "count": 120,
        "mean": 357.1666666666667,
        "median": 343.5,
        "p95": 575,
        "maximum": 999
      },
      "thinking_tokens": {
        "count": 120,
        "mean": 0,
        "median": 0.0,
        "p95": 0,
        "maximum": 0
      },
      "server_tokens_per_second": {
        "count": 120,
        "mean": 206.93643655442708,
        "median": 207.052891701266,
        "p95": 209.02668301512233,
        "maximum": 210.65130791776687
      },
      "first_token_seconds": {
        "count": 120,
        "mean": 0.06213989166666667,
        "median": 0.057375499999999996,
        "p95": 0.104512,
        "maximum": 0.159791
      },
      "no_final": 0
    },
    {
      "configuration": "gemma4-e2b/on",
      "condition": "no_source",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 28.825,
        "fully_correct": 10,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.245,
        "answers_with_material_errors": 107,
        "material_error_assessed": 120,
        "pending_reviews": 1,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 960.3166666666667,
        "mean_client_seconds": 4.707328449542304
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 4.707328449542304,
        "median_seconds": 4.522743173496565,
        "p90_seconds": 5.958411746949423,
        "p95_seconds": 6.925639235007111,
        "p99_seconds": 8.067634281003848,
        "max_seconds": 10.996482214017306,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 4.522743173496565,
        "final_returned_p95_seconds": 6.925639235007111,
        "first_token_timing_cases": 120,
        "first_token_p95_seconds": 0.047621000000000004,
        "observed_above_seconds": {
          "120": {
            "count": 0,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 3.9347740469966084,
        "p75_seconds": 5.372798163967673,
        "final_answer_completion": {
          "count": 120,
          "mean": 4.707328449542304,
          "median": 4.522743173496565,
          "p95": 6.925639235007111,
          "maximum": 10.996482214017306,
          "p25": 3.9347740469966084,
          "p75": 5.372798163967673
        },
        "above_30_seconds": 0,
        "bins": [
          82,
          37,
          1,
          0,
          0
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 960.3166666666667,
        "median": 924.0,
        "p95": 1421,
        "maximum": 2245
      },
      "input_tokens": {
        "count": 120,
        "mean": 124.33333333333333,
        "median": 116.0,
        "p95": 183,
        "maximum": 294
      },
      "thinking_tokens": {
        "count": 120,
        "mean": 675.5166666666667,
        "median": 650.0,
        "p95": 955,
        "maximum": 1495
      },
      "server_tokens_per_second": {
        "count": 120,
        "mean": 206.87896469619108,
        "median": 207.06801845829352,
        "p95": 207.7996237615422,
        "maximum": 208.43860687828126
      },
      "first_token_seconds": {
        "count": 120,
        "mean": 0.042842958333333334,
        "median": 0.040534,
        "p95": 0.047621000000000004,
        "maximum": 0.131977
      },
      "no_final": 0
    },
    {
      "configuration": "gemma4-e2b/on",
      "condition": "supplied_passage",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 86.84166666666667,
        "fully_correct": 76,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.8352777777777778,
        "answers_with_material_errors": 33,
        "material_error_assessed": 120,
        "pending_reviews": 0,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 837.1916666666667,
        "mean_client_seconds": 4.139805731367475
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 4.139805731367475,
        "median_seconds": 3.8559084575099405,
        "p90_seconds": 5.956665334000718,
        "p95_seconds": 6.569992751989048,
        "p99_seconds": 8.511655547015835,
        "max_seconds": 10.442087322997395,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 3.8559084575099405,
        "final_returned_p95_seconds": 6.569992751989048,
        "first_token_timing_cases": 120,
        "first_token_p95_seconds": 0.100721,
        "observed_above_seconds": {
          "120": {
            "count": 0,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 3.1147124849958345,
        "p75_seconds": 4.667062469001394,
        "final_answer_completion": {
          "count": 120,
          "mean": 4.139805731367475,
          "median": 3.8559084575099405,
          "p95": 6.569992751989048,
          "maximum": 10.442087322997395,
          "p25": 3.1147124849958345,
          "p75": 4.667062469001394
        },
        "above_30_seconds": 0,
        "bins": [
          95,
          24,
          1,
          0,
          0
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 837.1916666666667,
        "median": 784.5,
        "p95": 1334,
        "maximum": 2090
      },
      "input_tokens": {
        "count": 120,
        "mean": 359.1666666666667,
        "median": 345.5,
        "p95": 577,
        "maximum": 1001
      },
      "thinking_tokens": {
        "count": 120,
        "mean": 573.875,
        "median": 540.0,
        "p95": 953,
        "maximum": 1488
      },
      "server_tokens_per_second": {
        "count": 120,
        "mean": 206.59109639725347,
        "median": 206.71082036677552,
        "p95": 207.63291649887796,
        "maximum": 208.10760814989032
      },
      "first_token_seconds": {
        "count": 120,
        "mean": 0.06271653333333334,
        "median": 0.060014,
        "p95": 0.100721,
        "maximum": 0.130471
      },
      "no_final": 0
    },
    {
      "configuration": "k2-horizon-3.7b/low",
      "condition": "no_source",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 22.1,
        "fully_correct": 1,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.17097222222222222,
        "answers_with_material_errors": 115,
        "material_error_assessed": 120,
        "pending_reviews": 7,
        "operationally_flagged_answers": 24,
        "output_ceiling_hits": 23,
        "thinking_marker_breaches": 0,
        "direct_failures": 1,
        "regenerated_answers": 0,
        "mean_output_tokens": 6853.35,
        "mean_client_seconds": 106.64435930660848
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 106.86878929514181,
        "median_seconds": 5.0726538874987455,
        "p90_seconds": 522.565193628001,
        "p95_seconds": 525.5132120510007,
        "p99_seconds": 539.4491788619998,
        "max_seconds": 555.1688007489993,
        "nonempty_finals": 119,
        "no_final_cases": 1,
        "output_ceiling_hits": 23,
        "retried_cases": 3,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 5.03245187399807,
        "final_returned_p95_seconds": 525.8819249480002,
        "first_token_timing_cases": 0,
        "first_token_p95_seconds": null,
        "observed_above_seconds": {
          "120": {
            "count": 24,
            "denominator": 120
          },
          "180": {
            "count": 23,
            "denominator": 120
          },
          "300": {
            "count": 23,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 2.9458580930004246,
        "p75_seconds": 12.673349598000641,
        "final_answer_completion": {
          "count": 119,
          "mean": 107.60143212683208,
          "median": 5.03245187399807,
          "p95": 525.8819249480002,
          "maximum": 555.1688007489993,
          "p25": 2.9458580930004246,
          "p75": 12.673349598000641
        },
        "above_30_seconds": 27,
        "bins": [
          57,
          27,
          8,
          1,
          27
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 6853.35,
        "median": 506.5,
        "p95": 32768,
        "maximum": 32768
      },
      "input_tokens": {
        "count": 120,
        "mean": 121.18333333333334,
        "median": 113.0,
        "p95": 179,
        "maximum": 288
      },
      "thinking_tokens": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "server_tokens_per_second": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "first_token_seconds": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "no_final": 1
    },
    {
      "configuration": "k2-horizon-3.7b/low",
      "condition": "supplied_passage",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 70.7,
        "fully_correct": 33,
        "profile_assessed": 119,
        "mean_fully_met_claim_coverage": 0.6630555555555555,
        "answers_with_material_errors": 81,
        "material_error_assessed": 118,
        "pending_reviews": 8,
        "operationally_flagged_answers": 6,
        "output_ceiling_hits": 6,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 2106.5583333333334,
        "mean_client_seconds": 31.19163628355847
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 31.19163628355847,
        "median_seconds": 3.2705849434996708,
        "p90_seconds": 8.163189139000679,
        "p95_seconds": 89.80997591900086,
        "p99_seconds": 528.7268088199999,
        "max_seconds": 529.2080322930015,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 6,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 3.2705849434996708,
        "final_returned_p95_seconds": 89.80997591900086,
        "first_token_timing_cases": 0,
        "first_token_p95_seconds": null,
        "observed_above_seconds": {
          "120": {
            "count": 6,
            "denominator": 120
          },
          "180": {
            "count": 6,
            "denominator": 120
          },
          "300": {
            "count": 6,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 2.1468084669977543,
        "p75_seconds": 4.6954603560006944,
        "final_answer_completion": {
          "count": 120,
          "mean": 31.19163628355847,
          "median": 3.2705849434996708,
          "p95": 89.80997591900086,
          "maximum": 529.2080322930015,
          "p25": 2.1468084669977543,
          "p75": 4.6954603560006944
        },
        "above_30_seconds": 8,
        "bins": [
          94,
          14,
          3,
          1,
          8
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 2106.5583333333334,
        "median": 288.0,
        "p95": 8904,
        "maximum": 32768
      },
      "input_tokens": {
        "count": 120,
        "mean": 328.4916666666667,
        "median": 318.5,
        "p95": 522,
        "maximum": 915
      },
      "thinking_tokens": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "server_tokens_per_second": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "first_token_seconds": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "no_final": 0
    },
    {
      "configuration": "granite4.2-3b/off",
      "condition": "no_source",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 33.858333333333334,
        "fully_correct": 2,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.27819444444444447,
        "answers_with_material_errors": 118,
        "material_error_assessed": 120,
        "pending_reviews": 3,
        "operationally_flagged_answers": 89,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 89,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 2866.6916666666666,
        "mean_client_seconds": 17.536824373217314
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 17.536824373217314,
        "median_seconds": 11.627608879498439,
        "p90_seconds": 35.80614589099423,
        "p95_seconds": 52.12359667501005,
        "p99_seconds": 97.73319554699992,
        "max_seconds": 100.92142155799957,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 11.627608879498439,
        "final_returned_p95_seconds": 52.12359667501005,
        "first_token_timing_cases": 0,
        "first_token_p95_seconds": null,
        "observed_above_seconds": {
          "120": {
            "count": 0,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 5.959842812997522,
        "p75_seconds": 20.273498154987465,
        "final_answer_completion": {
          "count": 120,
          "mean": 17.536824373217314,
          "median": 11.627608879498439,
          "p95": 52.12359667501005,
          "maximum": 100.92142155799957,
          "p25": 5.959842812997522,
          "p75": 20.273498154987465
        },
        "above_30_seconds": 22,
        "bins": [
          25,
          26,
          37,
          10,
          22
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 2866.6916666666666,
        "median": 1994.0,
        "p95": 8270,
        "maximum": 15282
      },
      "input_tokens": {
        "count": 120,
        "mean": 123.20833333333333,
        "median": 115.0,
        "p95": 181,
        "maximum": 295
      },
      "thinking_tokens": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "server_tokens_per_second": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "first_token_seconds": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "no_final": 0
    },
    {
      "configuration": "granite4.2-3b/off",
      "condition": "supplied_passage",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 84.55833333333334,
        "fully_correct": 49,
        "profile_assessed": 119,
        "mean_fully_met_claim_coverage": 0.8165277777777777,
        "answers_with_material_errors": 69,
        "material_error_assessed": 119,
        "pending_reviews": 1,
        "operationally_flagged_answers": 67,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 67,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 1832.55,
        "mean_client_seconds": 11.2999894266919
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 11.2999894266919,
        "median_seconds": 6.8670859089979785,
        "p90_seconds": 19.910446284004138,
        "p95_seconds": 37.386897885997314,
        "p99_seconds": 83.05889781400037,
        "max_seconds": 135.73921225199592,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 6.8670859089979785,
        "final_returned_p95_seconds": 37.386897885997314,
        "first_token_timing_cases": 0,
        "first_token_p95_seconds": null,
        "observed_above_seconds": {
          "120": {
            "count": 1,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 2.4767175600136397,
        "p75_seconds": 12.979480125999544,
        "final_answer_completion": {
          "count": 120,
          "mean": 11.2999894266919,
          "median": 6.8670859089979785,
          "p95": 37.386897885997314,
          "maximum": 135.73921225199592,
          "p25": 2.4767175600136397,
          "p75": 12.979480125999544
        },
        "above_30_seconds": 8,
        "bins": [
          51,
          23,
          34,
          4,
          8
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 1832.55,
        "median": 1157.0,
        "p95": 6097,
        "maximum": 19672
      },
      "input_tokens": {
        "count": 120,
        "mean": 358.39166666666665,
        "median": 346.5,
        "p95": 578,
        "maximum": 1005
      },
      "thinking_tokens": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "server_tokens_per_second": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "first_token_seconds": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "no_final": 0
    },
    {
      "configuration": "granite4.2-3b/on",
      "condition": "no_source",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 38.88333333333333,
        "fully_correct": 9,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.33319444444444446,
        "answers_with_material_errors": 111,
        "material_error_assessed": 120,
        "pending_reviews": 1,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 5508.375,
        "mean_client_seconds": 34.57511501556607
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 34.57511501556607,
        "median_seconds": 26.374946923504467,
        "p90_seconds": 65.69577317099902,
        "p95_seconds": 86.68572923299507,
        "p99_seconds": 116.41093123900646,
        "max_seconds": 204.30458438499772,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 26.374946923504467,
        "final_returned_p95_seconds": 86.68572923299507,
        "first_token_timing_cases": 0,
        "first_token_p95_seconds": null,
        "observed_above_seconds": {
          "120": {
            "count": 1,
            "denominator": 120
          },
          "180": {
            "count": 1,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 15.907932058005827,
        "p75_seconds": 44.13561461601057,
        "final_answer_completion": {
          "count": 120,
          "mean": 34.57511501556607,
          "median": 26.374946923504467,
          "p95": 86.68572923299507,
          "maximum": 204.30458438499772,
          "p25": 15.907932058005827,
          "p75": 44.13561461601057
        },
        "above_30_seconds": 54,
        "bins": [
          1,
          11,
          29,
          25,
          54
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 5508.375,
        "median": 4377.0,
        "p95": 13574,
        "maximum": 28099
      },
      "input_tokens": {
        "count": 120,
        "mean": 123.20833333333333,
        "median": 115.0,
        "p95": 181,
        "maximum": 295
      },
      "thinking_tokens": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "server_tokens_per_second": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "first_token_seconds": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "no_final": 0
    },
    {
      "configuration": "granite4.2-3b/on",
      "condition": "supplied_passage",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 90.58333333333333,
        "fully_correct": 76,
        "profile_assessed": 119,
        "mean_fully_met_claim_coverage": 0.8825,
        "answers_with_material_errors": 43,
        "material_error_assessed": 119,
        "pending_reviews": 1,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 2508.1916666666666,
        "mean_client_seconds": 15.560693055508697
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 15.560693055508697,
        "median_seconds": 8.51492815700476,
        "p90_seconds": 29.18252176900569,
        "p95_seconds": 45.17118737300916,
        "p99_seconds": 94.05363320399192,
        "max_seconds": 124.34756956799538,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 8.51492815700476,
        "final_returned_p95_seconds": 45.17118737300916,
        "first_token_timing_cases": 0,
        "first_token_p95_seconds": null,
        "observed_above_seconds": {
          "120": {
            "count": 1,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 5.164923757998622,
        "p75_seconds": 18.353012742998544,
        "final_answer_completion": {
          "count": 120,
          "mean": 15.560693055508697,
          "median": 8.51492815700476,
          "p95": 45.17118737300916,
          "maximum": 124.34756956799538,
          "p25": 5.164923757998622,
          "p75": 18.353012742998544
        },
        "above_30_seconds": 12,
        "bins": [
          27,
          43,
          25,
          13,
          12
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 2508.1916666666666,
        "median": 1443.5,
        "p95": 7364,
        "maximum": 18336
      },
      "input_tokens": {
        "count": 120,
        "mean": 358.39166666666665,
        "median": 346.5,
        "p95": 578,
        "maximum": 1005
      },
      "thinking_tokens": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "server_tokens_per_second": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "first_token_seconds": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "no_final": 0
    },
    {
      "configuration": "qwen3.5-4b/off",
      "condition": "no_source",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 36.69166666666667,
        "fully_correct": 0,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.30569444444444444,
        "answers_with_material_errors": 120,
        "material_error_assessed": 120,
        "pending_reviews": 23,
        "operationally_flagged_answers": 29,
        "output_ceiling_hits": 1,
        "thinking_marker_breaches": 28,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 2916.616666666667,
        "mean_client_seconds": 22.127485552891933
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 22.127485552891933,
        "median_seconds": 6.114862078502483,
        "p90_seconds": 66.43627836499945,
        "p95_seconds": 82.86190171299677,
        "p99_seconds": 107.71195344100124,
        "max_seconds": 300.5635435330041,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 1,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 6.114862078502483,
        "final_returned_p95_seconds": 82.86190171299677,
        "first_token_timing_cases": 0,
        "first_token_p95_seconds": null,
        "observed_above_seconds": {
          "120": {
            "count": 1,
            "denominator": 120
          },
          "180": {
            "count": 1,
            "denominator": 120
          },
          "300": {
            "count": 1,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 4.263538664003136,
        "p75_seconds": 23.148665054992307,
        "final_answer_completion": {
          "count": 120,
          "mean": 22.127485552891933,
          "median": 6.114862078502483,
          "p95": 82.86190171299677,
          "maximum": 300.5635435330041,
          "p25": 4.263538664003136,
          "p75": 23.148665054992307
        },
        "above_30_seconds": 30,
        "bins": [
          43,
          37,
          8,
          2,
          30
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 2916.616666666667,
        "median": 834.0,
        "p95": 11057,
        "maximum": 32768
      },
      "input_tokens": {
        "count": 120,
        "mean": 124.375,
        "median": 116.0,
        "p95": 182,
        "maximum": 297
      },
      "thinking_tokens": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "server_tokens_per_second": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "first_token_seconds": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "no_final": 0
    },
    {
      "configuration": "qwen3.5-4b/off",
      "condition": "supplied_passage",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 87.36666666666666,
        "fully_correct": 51,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.8476388888888889,
        "answers_with_material_errors": 68,
        "material_error_assessed": 120,
        "pending_reviews": 5,
        "operationally_flagged_answers": 6,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 6,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 986.5,
        "mean_client_seconds": 7.438834479966075
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 7.438834479966075,
        "median_seconds": 3.5740008395077894,
        "p90_seconds": 8.9167451300018,
        "p95_seconds": 16.167494409004576,
        "p99_seconds": 79.12350526599039,
        "max_seconds": 101.28867584100226,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 3.5740008395077894,
        "final_returned_p95_seconds": 16.167494409004576,
        "first_token_timing_cases": 0,
        "first_token_p95_seconds": null,
        "observed_above_seconds": {
          "120": {
            "count": 0,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 2.5554613540007267,
        "p75_seconds": 5.134720807996928,
        "final_answer_completion": {
          "count": 120,
          "mean": 7.438834479966075,
          "median": 3.5740008395077894,
          "p95": 16.167494409004576,
          "maximum": 101.28867584100226,
          "p25": 2.5554613540007267,
          "p75": 5.134720807996928
        },
        "above_30_seconds": 6,
        "bins": [
          88,
          22,
          4,
          0,
          6
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 986.5,
        "median": 467.0,
        "p95": 2222,
        "maximum": 13376
      },
      "input_tokens": {
        "count": 120,
        "mean": 349.78333333333336,
        "median": 337.5,
        "p95": 556,
        "maximum": 964
      },
      "thinking_tokens": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "server_tokens_per_second": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "first_token_seconds": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "no_final": 0
    },
    {
      "configuration": "qwen3.5-4b/on",
      "condition": "no_source",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 40.59166666666667,
        "fully_correct": 2,
        "profile_assessed": 119,
        "mean_fully_met_claim_coverage": 0.34555555555555556,
        "answers_with_material_errors": 116,
        "material_error_assessed": 119,
        "pending_reviews": 18,
        "operationally_flagged_answers": 1,
        "output_ceiling_hits": 1,
        "thinking_marker_breaches": 0,
        "direct_failures": 1,
        "regenerated_answers": 0,
        "mean_output_tokens": 11715.916666666666,
        "mean_client_seconds": 90.43474560724935
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 90.43474560724935,
        "median_seconds": 83.77937390448642,
        "p90_seconds": 138.2841383019986,
        "p95_seconds": 163.26857816400297,
        "p99_seconds": 193.32805091701448,
        "max_seconds": 301.8480701780063,
        "nonempty_finals": 119,
        "no_final_cases": 1,
        "output_ceiling_hits": 1,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 83.75592601599055,
        "final_returned_p95_seconds": 163.26857816400297,
        "first_token_timing_cases": 0,
        "first_token_p95_seconds": null,
        "observed_above_seconds": {
          "120": {
            "count": 22,
            "denominator": 120
          },
          "180": {
            "count": 3,
            "denominator": 120
          },
          "300": {
            "count": 1,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 65.06239960300445,
        "p75_seconds": 105.07801196399669,
        "final_answer_completion": {
          "count": 119,
          "mean": 88.65816304783122,
          "median": 83.75592601599055,
          "p95": 163.26857816400297,
          "maximum": 193.32805091701448,
          "p25": 65.06239960300445,
          "p75": 105.07801196399669
        },
        "above_30_seconds": 117,
        "bins": [
          0,
          0,
          1,
          2,
          117
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 11715.916666666666,
        "median": 10995.5,
        "p95": 19566,
        "maximum": 32768
      },
      "input_tokens": {
        "count": 120,
        "mean": 122.375,
        "median": 114.0,
        "p95": 180,
        "maximum": 295
      },
      "thinking_tokens": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "server_tokens_per_second": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "first_token_seconds": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "no_final": 1
    },
    {
      "configuration": "qwen3.5-4b/on",
      "condition": "supplied_passage",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 95.38333333333334,
        "fully_correct": 90,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.9331944444444444,
        "answers_with_material_errors": 28,
        "material_error_assessed": 120,
        "pending_reviews": 1,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 5193.15,
        "mean_client_seconds": 38.75811221433311
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 38.75811221433311,
        "median_seconds": 32.70316506900417,
        "p90_seconds": 74.03098855999997,
        "p95_seconds": 94.70089313600329,
        "p99_seconds": 148.31651260999206,
        "max_seconds": 166.62847473600414,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 32.70316506900417,
        "final_returned_p95_seconds": 94.70089313600329,
        "first_token_timing_cases": 0,
        "first_token_p95_seconds": null,
        "observed_above_seconds": {
          "120": {
            "count": 3,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 17.909592307987623,
        "p75_seconds": 47.11463881999953,
        "final_answer_completion": {
          "count": 120,
          "mean": 38.75811221433311,
          "median": 32.70316506900417,
          "p95": 94.70089313600329,
          "maximum": 166.62847473600414,
          "p25": 17.909592307987623,
          "p75": 47.11463881999953
        },
        "above_30_seconds": 67,
        "bins": [
          0,
          8,
          25,
          20,
          67
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 5193.15,
        "median": 4501.0,
        "p95": 12534,
        "maximum": 20290
      },
      "input_tokens": {
        "count": 120,
        "mean": 347.78333333333336,
        "median": 335.5,
        "p95": 554,
        "maximum": 962
      },
      "thinking_tokens": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "server_tokens_per_second": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "first_token_seconds": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "no_final": 0
    },
    {
      "configuration": "granite4.2-8b/on",
      "condition": "no_source",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 44.575,
        "fully_correct": 2,
        "profile_assessed": 119,
        "mean_fully_met_claim_coverage": 0.37958333333333333,
        "answers_with_material_errors": 117,
        "material_error_assessed": 119,
        "pending_reviews": 4,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 4734.825,
        "mean_client_seconds": 48.04009623179081
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 48.04009623179081,
        "median_seconds": 37.739013958504074,
        "p90_seconds": 85.81496201298432,
        "p95_seconds": 96.41692897598841,
        "p99_seconds": 202.17382362298667,
        "max_seconds": 219.2593050459982,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 37.739013958504074,
        "final_returned_p95_seconds": 96.41692897598841,
        "first_token_timing_cases": 0,
        "first_token_p95_seconds": null,
        "observed_above_seconds": {
          "120": {
            "count": 3,
            "denominator": 120
          },
          "180": {
            "count": 2,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 25.79880037700059,
        "p75_seconds": 64.16362267101067,
        "final_answer_completion": {
          "count": 120,
          "mean": 48.04009623179081,
          "median": 37.739013958504074,
          "p95": 96.41692897598841,
          "maximum": 219.2593050459982,
          "p25": 25.79880037700059,
          "p75": 64.16362267101067
        },
        "above_30_seconds": 81,
        "bins": [
          0,
          0,
          21,
          18,
          81
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 4734.825,
        "median": 3825.0,
        "p95": 9269,
        "maximum": 19375
      },
      "input_tokens": {
        "count": 120,
        "mean": 123.20833333333333,
        "median": 115.0,
        "p95": 181,
        "maximum": 295
      },
      "thinking_tokens": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "server_tokens_per_second": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "first_token_seconds": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "no_final": 0
    },
    {
      "configuration": "granite4.2-8b/on",
      "condition": "supplied_passage",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 93.96666666666667,
        "fully_correct": 63,
        "profile_assessed": 117,
        "mean_fully_met_claim_coverage": 0.9215277777777777,
        "answers_with_material_errors": 54,
        "material_error_assessed": 117,
        "pending_reviews": 4,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 2971.891666666667,
        "mean_client_seconds": 30.398619683539923
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 30.398619683539923,
        "median_seconds": 18.47347018450091,
        "p90_seconds": 57.82827217900194,
        "p95_seconds": 97.1606799820147,
        "p99_seconds": 153.6355670490011,
        "max_seconds": 304.3662539739744,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 18.47347018450091,
        "final_returned_p95_seconds": 97.1606799820147,
        "first_token_timing_cases": 0,
        "first_token_p95_seconds": null,
        "observed_above_seconds": {
          "120": {
            "count": 4,
            "denominator": 120
          },
          "180": {
            "count": 1,
            "denominator": 120
          },
          "300": {
            "count": 1,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 13.288219420996029,
        "p75_seconds": 30.699411610985408,
        "final_answer_completion": {
          "count": 120,
          "mean": 30.398619683539923,
          "median": 18.47347018450091,
          "p95": 97.1606799820147,
          "maximum": 304.3662539739744,
          "p25": 13.288219420996029,
          "p75": 30.699411610985408
        },
        "above_30_seconds": 31,
        "bins": [
          0,
          13,
          55,
          21,
          31
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 2971.891666666667,
        "median": 1884.0,
        "p95": 9408,
        "maximum": 25497
      },
      "input_tokens": {
        "count": 120,
        "mean": 358.39166666666665,
        "median": 346.5,
        "p95": 578,
        "maximum": 1005
      },
      "thinking_tokens": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "server_tokens_per_second": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "first_token_seconds": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "no_final": 0
    },
    {
      "configuration": "granite4.2-8b/off",
      "condition": "no_source",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 41.09166666666667,
        "fully_correct": 2,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.3473611111111111,
        "answers_with_material_errors": 118,
        "material_error_assessed": 120,
        "pending_reviews": 4,
        "operationally_flagged_answers": 78,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 78,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 3264.0833333333335,
        "mean_client_seconds": 33.18850703368516
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 33.18850703368516,
        "median_seconds": 22.379449365500477,
        "p90_seconds": 62.04956451698672,
        "p95_seconds": 86.87513911200222,
        "p99_seconds": 146.75476339098532,
        "max_seconds": 326.8716572630219,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 22.379449365500477,
        "final_returned_p95_seconds": 86.87513911200222,
        "first_token_timing_cases": 0,
        "first_token_p95_seconds": null,
        "observed_above_seconds": {
          "120": {
            "count": 4,
            "denominator": 120
          },
          "180": {
            "count": 1,
            "denominator": 120
          },
          "300": {
            "count": 1,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 13.03906223201193,
        "p75_seconds": 37.6958039989986,
        "final_answer_completion": {
          "count": 120,
          "mean": 33.18850703368516,
          "median": 22.379449365500477,
          "p95": 86.87513911200222,
          "maximum": 326.8716572630219,
          "p25": 13.03906223201193,
          "p75": 37.6958039989986
        },
        "above_30_seconds": 46,
        "bins": [
          0,
          14,
          40,
          20,
          46
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 3264.0833333333335,
        "median": 2285.5,
        "p95": 8363,
        "maximum": 27161
      },
      "input_tokens": {
        "count": 120,
        "mean": 123.20833333333333,
        "median": 115.0,
        "p95": 181,
        "maximum": 295
      },
      "thinking_tokens": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "server_tokens_per_second": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "first_token_seconds": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "no_final": 0
    },
    {
      "configuration": "granite4.2-8b/off",
      "condition": "supplied_passage",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 88.825,
        "fully_correct": 63,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.8708333333333333,
        "answers_with_material_errors": 57,
        "material_error_assessed": 120,
        "pending_reviews": 1,
        "operationally_flagged_answers": 61,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 61,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 1685.0833333333333,
        "mean_client_seconds": 16.949836852439088
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 16.949836852439088,
        "median_seconds": 9.41112702449027,
        "p90_seconds": 33.4469473589852,
        "p95_seconds": 46.68501166198985,
        "p99_seconds": 94.5617580129765,
        "max_seconds": 152.06399614401744,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 9.41112702449027,
        "final_returned_p95_seconds": 46.68501166198985,
        "first_token_timing_cases": 0,
        "first_token_p95_seconds": null,
        "observed_above_seconds": {
          "120": {
            "count": 1,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 6.145904293021886,
        "p75_seconds": 19.536974838003516,
        "final_answer_completion": {
          "count": 120,
          "mean": 16.949836852439088,
          "median": 9.41112702449027,
          "p95": 46.68501166198985,
          "maximum": 152.06399614401744,
          "p25": 6.145904293021886,
          "p75": 19.536974838003516
        },
        "above_30_seconds": 18,
        "bins": [
          17,
          44,
          31,
          10,
          18
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 1685.0833333333333,
        "median": 950.0,
        "p95": 4691,
        "maximum": 13886
      },
      "input_tokens": {
        "count": 120,
        "mean": 358.39166666666665,
        "median": 346.5,
        "p95": 578,
        "maximum": 1005
      },
      "thinking_tokens": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "server_tokens_per_second": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "first_token_seconds": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "no_final": 0
    },
    {
      "configuration": "lfm2.5-8b-a1b/native",
      "condition": "no_source",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 27.625,
        "fully_correct": 0,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.21611111111111111,
        "answers_with_material_errors": 119,
        "material_error_assessed": 120,
        "pending_reviews": 2,
        "operationally_flagged_answers": 1,
        "output_ceiling_hits": 1,
        "thinking_marker_breaches": 0,
        "direct_failures": 1,
        "regenerated_answers": 0,
        "mean_output_tokens": 4046.925,
        "mean_client_seconds": 13.336282462099673
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 13.336282462099673,
        "median_seconds": 9.907918013996095,
        "p90_seconds": 22.2773051999975,
        "p95_seconds": 29.351412558025913,
        "p99_seconds": 56.42283325601602,
        "max_seconds": 112.16383835699526,
        "nonempty_finals": 119,
        "no_final_cases": 1,
        "output_ceiling_hits": 1,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 9.788082423998276,
        "final_returned_p95_seconds": 29.351412558025913,
        "first_token_timing_cases": 120,
        "first_token_p95_seconds": 0.116819,
        "observed_above_seconds": {
          "120": {
            "count": 0,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 6.383746783976676,
        "p75_seconds": 17.538372928014724,
        "final_answer_completion": {
          "count": 119,
          "mean": 12.505798799117358,
          "median": 9.788082423998276,
          "p95": 29.351412558025913,
          "maximum": 56.42283325601602,
          "p25": 6.383746783976676,
          "p75": 17.538372928014724
        },
        "above_30_seconds": 5,
        "bins": [
          14,
          46,
          40,
          15,
          5
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 4046.925,
        "median": 3037.5,
        "p95": 8904,
        "maximum": 32768
      },
      "input_tokens": {
        "count": 120,
        "mean": 124.6,
        "median": 116.0,
        "p95": 181,
        "maximum": 301
      },
      "thinking_tokens": {
        "count": 120,
        "mean": 3656.1,
        "median": 2584.0,
        "p95": 8328,
        "maximum": 32766
      },
      "server_tokens_per_second": {
        "count": 120,
        "mean": 311.49877122227093,
        "median": 310.5096004097762,
        "p95": 321.0085772548855,
        "maximum": 328.9777420320027
      },
      "first_token_seconds": {
        "count": 120,
        "mean": 0.05459571666666667,
        "median": 0.042054499999999995,
        "p95": 0.116819,
        "maximum": 0.152178
      },
      "no_final": 1
    },
    {
      "configuration": "lfm2.5-8b-a1b/native",
      "condition": "supplied_passage",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 71.40833333333333,
        "fully_correct": 32,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.6590277777777778,
        "answers_with_material_errors": 86,
        "material_error_assessed": 120,
        "pending_reviews": 3,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 1729.875,
        "mean_client_seconds": 5.723916339183051
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 5.723916339183051,
        "median_seconds": 4.069626017007977,
        "p90_seconds": 10.44233370697475,
        "p95_seconds": 14.343253933999222,
        "p99_seconds": 22.32033113099169,
        "max_seconds": 46.16163652201067,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 4.069626017007977,
        "final_returned_p95_seconds": 14.343253933999222,
        "first_token_timing_cases": 120,
        "first_token_p95_seconds": 0.128366,
        "observed_above_seconds": {
          "120": {
            "count": 0,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 2.97586185898399,
        "p75_seconds": 5.651144048984861,
        "final_answer_completion": {
          "count": 120,
          "mean": 5.723916339183051,
          "median": 4.069626017007977,
          "p95": 14.343253933999222,
          "maximum": 46.16163652201067,
          "p25": 2.97586185898399,
          "p75": 5.651144048984861
        },
        "above_30_seconds": 1,
        "bins": [
          73,
          33,
          10,
          3,
          1
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 1729.875,
        "median": 1235.5,
        "p95": 4414,
        "maximum": 14076
      },
      "input_tokens": {
        "count": 120,
        "mean": 342.3833333333333,
        "median": 332.0,
        "p95": 546,
        "maximum": 961
      },
      "thinking_tokens": {
        "count": 120,
        "mean": 1368.6833333333334,
        "median": 852.5,
        "p95": 3867,
        "maximum": 13773
      },
      "server_tokens_per_second": {
        "count": 120,
        "mean": 313.52707575931635,
        "median": 314.4757126761283,
        "p95": 320.8534097212549,
        "maximum": 324.18276172159875
      },
      "first_token_seconds": {
        "count": 120,
        "mean": 0.06747885,
        "median": 0.053706000000000004,
        "p95": 0.128366,
        "maximum": 0.181173
      },
      "no_final": 0
    },
    {
      "configuration": "qwen3.5-9b/off",
      "condition": "no_source",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 51.88333333333333,
        "fully_correct": 5,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.45944444444444443,
        "answers_with_material_errors": 115,
        "material_error_assessed": 120,
        "pending_reviews": 10,
        "operationally_flagged_answers": 13,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 13,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 1701.375,
        "mean_client_seconds": 16.72947611001776
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 16.72947611001776,
        "median_seconds": 7.620644941518549,
        "p90_seconds": 52.43825192100485,
        "p95_seconds": 76.44514443399385,
        "p99_seconds": 107.81710012498661,
        "max_seconds": 125.00687698199181,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 7.620644941518549,
        "final_returned_p95_seconds": 76.44514443399385,
        "first_token_timing_cases": 0,
        "first_token_p95_seconds": null,
        "observed_above_seconds": {
          "120": {
            "count": 1,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 5.218570884026121,
        "p75_seconds": 11.776204310008325,
        "final_answer_completion": {
          "count": 120,
          "mean": 16.72947611001776,
          "median": 7.620644941518549,
          "p95": 76.44514443399385,
          "maximum": 125.00687698199181,
          "p25": 5.218570884026121,
          "p75": 11.776204310008325
        },
        "above_30_seconds": 17,
        "bins": [
          23,
          62,
          14,
          4,
          17
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 1701.375,
        "median": 766.5,
        "p95": 7835,
        "maximum": 12512
      },
      "input_tokens": {
        "count": 120,
        "mean": 124.375,
        "median": 116.0,
        "p95": 182,
        "maximum": 297
      },
      "thinking_tokens": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "server_tokens_per_second": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "first_token_seconds": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "no_final": 0
    },
    {
      "configuration": "qwen3.5-9b/off",
      "condition": "supplied_passage",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 89.24166666666666,
        "fully_correct": 49,
        "profile_assessed": 116,
        "mean_fully_met_claim_coverage": 0.8701388888888889,
        "answers_with_material_errors": 63,
        "material_error_assessed": 116,
        "pending_reviews": 9,
        "operationally_flagged_answers": 3,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 3,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 749.725,
        "mean_client_seconds": 7.4671080245427826
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 7.4671080245427826,
        "median_seconds": 4.855398231986328,
        "p90_seconds": 13.177598420996219,
        "p95_seconds": 19.39523829999962,
        "p99_seconds": 60.62675652699545,
        "max_seconds": 62.68317884899443,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 4.855398231986328,
        "final_returned_p95_seconds": 19.39523829999962,
        "first_token_timing_cases": 0,
        "first_token_p95_seconds": null,
        "observed_above_seconds": {
          "120": {
            "count": 0,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 3.7374641920032445,
        "p75_seconds": 6.772200307023013,
        "final_answer_completion": {
          "count": 120,
          "mean": 7.4671080245427826,
          "median": 4.855398231986328,
          "p95": 19.39523829999962,
          "maximum": 62.68317884899443,
          "p25": 3.7374641920032445,
          "p75": 6.772200307023013
        },
        "above_30_seconds": 4,
        "bins": [
          62,
          43,
          10,
          1,
          4
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 749.725,
        "median": 486.0,
        "p95": 1979,
        "maximum": 6445
      },
      "input_tokens": {
        "count": 120,
        "mean": 349.78333333333336,
        "median": 337.5,
        "p95": 556,
        "maximum": 964
      },
      "thinking_tokens": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "server_tokens_per_second": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "first_token_seconds": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "no_final": 0
    },
    {
      "configuration": "qwen3.5-9b/on",
      "condition": "no_source",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 45.3,
        "fully_correct": 8,
        "profile_assessed": 119,
        "mean_fully_met_claim_coverage": 0.4038888888888889,
        "answers_with_material_errors": 110,
        "material_error_assessed": 119,
        "pending_reviews": 22,
        "operationally_flagged_answers": 2,
        "output_ceiling_hits": 2,
        "thinking_marker_breaches": 0,
        "direct_failures": 1,
        "regenerated_answers": 0,
        "mean_output_tokens": 9077.375,
        "mean_client_seconds": 89.8962851640749
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 89.8962851640749,
        "median_seconds": 83.022500837993,
        "p90_seconds": 138.56031002098462,
        "p95_seconds": 158.42903493501944,
        "p99_seconds": 339.06330648000585,
        "max_seconds": 339.0751682180271,
        "nonempty_finals": 119,
        "no_final_cases": 1,
        "output_ceiling_hits": 2,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 82.89199672199902,
        "final_returned_p95_seconds": 158.42903493501944,
        "first_token_timing_cases": 0,
        "first_token_p95_seconds": null,
        "observed_above_seconds": {
          "120": {
            "count": 20,
            "denominator": 120
          },
          "180": {
            "count": 2,
            "denominator": 120
          },
          "300": {
            "count": 2,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 67.03282110000146,
        "p75_seconds": 98.99658961599926,
        "final_answer_completion": {
          "count": 119,
          "mean": 87.80244464881497,
          "median": 82.89199672199902,
          "p95": 158.42903493501944,
          "maximum": 339.0751682180271,
          "p25": 67.03282110000146,
          "p75": 98.99658961599926
        },
        "above_30_seconds": 116,
        "bins": [
          0,
          1,
          1,
          2,
          116
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 9077.375,
        "median": 8419.0,
        "p95": 15804,
        "maximum": 32768
      },
      "input_tokens": {
        "count": 120,
        "mean": 122.375,
        "median": 114.0,
        "p95": 180,
        "maximum": 295
      },
      "thinking_tokens": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "server_tokens_per_second": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "first_token_seconds": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "no_final": 1
    },
    {
      "configuration": "qwen3.5-9b/on",
      "condition": "supplied_passage",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 96.625,
        "fully_correct": 99,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.9554166666666667,
        "answers_with_material_errors": 19,
        "material_error_assessed": 120,
        "pending_reviews": 0,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 3577.925,
        "mean_client_seconds": 35.111306433840944
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 35.111306433840944,
        "median_seconds": 26.66372540150769,
        "p90_seconds": 66.45543938499759,
        "p95_seconds": 77.80049887401401,
        "p99_seconds": 138.38800971899764,
        "max_seconds": 145.64245316997403,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 26.66372540150769,
        "final_returned_p95_seconds": 77.80049887401401,
        "first_token_timing_cases": 0,
        "first_token_p95_seconds": null,
        "observed_above_seconds": {
          "120": {
            "count": 2,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 19.138037466997048,
        "p75_seconds": 40.515594926982885,
        "final_answer_completion": {
          "count": 120,
          "mean": 35.111306433840944,
          "median": 26.66372540150769,
          "p95": 77.80049887401401,
          "maximum": 145.64245316997403,
          "p25": 19.138037466997048,
          "p75": 40.515594926982885
        },
        "above_30_seconds": 53,
        "bins": [
          0,
          2,
          32,
          33,
          53
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 3577.925,
        "median": 2712.5,
        "p95": 7852,
        "maximum": 14684
      },
      "input_tokens": {
        "count": 120,
        "mean": 347.78333333333336,
        "median": 335.5,
        "p95": 554,
        "maximum": 962
      },
      "thinking_tokens": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "server_tokens_per_second": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "first_token_seconds": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "no_final": 0
    },
    {
      "configuration": "granite4.2-30b/on",
      "condition": "no_source",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 45.55,
        "fully_correct": 1,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.4122222222222222,
        "answers_with_material_errors": 119,
        "material_error_assessed": 120,
        "pending_reviews": 12,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 2336.325,
        "mean_client_seconds": 102.97626134025
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 102.97626134025,
        "median_seconds": 71.84132410149323,
        "p90_seconds": 179.39653907099273,
        "p95_seconds": 239.7329822370084,
        "p99_seconds": 618.5921809279826,
        "max_seconds": 794.8344104519929,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 71.84132410149323,
        "final_returned_p95_seconds": 239.7329822370084,
        "first_token_timing_cases": 0,
        "first_token_p95_seconds": null,
        "observed_above_seconds": {
          "120": {
            "count": 22,
            "denominator": 120
          },
          "180": {
            "count": 12,
            "denominator": 120
          },
          "300": {
            "count": 4,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 55.852897441975074,
        "p75_seconds": 97.83876022600452,
        "final_answer_completion": {
          "count": 120,
          "mean": 102.97626134025,
          "median": 71.84132410149323,
          "p95": 239.7329822370084,
          "maximum": 794.8344104519929,
          "p25": 55.852897441975074,
          "p75": 97.83876022600452
        },
        "above_30_seconds": 119,
        "bins": [
          0,
          0,
          0,
          1,
          119
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 2336.325,
        "median": 1860.5,
        "p95": 4890,
        "maximum": 11321
      },
      "input_tokens": {
        "count": 120,
        "mean": 123.20833333333333,
        "median": 115.0,
        "p95": 181,
        "maximum": 295
      },
      "thinking_tokens": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "server_tokens_per_second": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "first_token_seconds": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "no_final": 0
    },
    {
      "configuration": "granite4.2-30b/on",
      "condition": "supplied_passage",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 95.91666666666667,
        "fully_correct": 69,
        "profile_assessed": 119,
        "mean_fully_met_claim_coverage": 0.9463888888888888,
        "answers_with_material_errors": 49,
        "material_error_assessed": 119,
        "pending_reviews": 3,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 1493.2916666666667,
        "mean_client_seconds": 64.74146516768451
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 64.74146516768451,
        "median_seconds": 52.041306866987725,
        "p90_seconds": 80.43559403499239,
        "p95_seconds": 97.29196825801046,
        "p99_seconds": 559.1574552730017,
        "max_seconds": 666.3046439729806,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 52.041306866987725,
        "final_returned_p95_seconds": 97.29196825801046,
        "first_token_timing_cases": 0,
        "first_token_p95_seconds": null,
        "observed_above_seconds": {
          "120": {
            "count": 3,
            "denominator": 120
          },
          "180": {
            "count": 2,
            "denominator": 120
          },
          "300": {
            "count": 2,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 43.510723234008765,
        "p75_seconds": 62.60568271597731,
        "final_answer_completion": {
          "count": 120,
          "mean": 64.74146516768451,
          "median": 52.041306866987725,
          "p95": 97.29196825801046,
          "maximum": 666.3046439729806,
          "p25": 43.510723234008765,
          "p75": 62.60568271597731
        },
        "above_30_seconds": 118,
        "bins": [
          0,
          0,
          0,
          2,
          118
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 1493.2916666666667,
        "median": 1310.5,
        "p95": 2243,
        "maximum": 9304
      },
      "input_tokens": {
        "count": 120,
        "mean": 358.39166666666665,
        "median": 346.5,
        "p95": 578,
        "maximum": 1005
      },
      "thinking_tokens": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "server_tokens_per_second": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "first_token_seconds": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "no_final": 0
    },
    {
      "configuration": "granite4.2-30b/off",
      "condition": "no_source",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 44.06666666666667,
        "fully_correct": 0,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.3595833333333333,
        "answers_with_material_errors": 120,
        "material_error_assessed": 120,
        "pending_reviews": 7,
        "operationally_flagged_answers": 1,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 1,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 1406.5666666666666,
        "mean_client_seconds": 60.31475436160836
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 60.31475436160836,
        "median_seconds": 40.138928431493696,
        "p90_seconds": 99.73870880898903,
        "p95_seconds": 116.4921896349988,
        "p99_seconds": 505.9988825870096,
        "max_seconds": 515.3523483290046,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 40.138928431493696,
        "final_returned_p95_seconds": 116.4921896349988,
        "first_token_timing_cases": 0,
        "first_token_p95_seconds": null,
        "observed_above_seconds": {
          "120": {
            "count": 6,
            "denominator": 120
          },
          "180": {
            "count": 4,
            "denominator": 120
          },
          "300": {
            "count": 4,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 26.867685119999805,
        "p75_seconds": 57.83864623098634,
        "final_answer_completion": {
          "count": 120,
          "mean": 60.31475436160836,
          "median": 40.138928431493696,
          "p95": 116.4921896349988,
          "maximum": 515.3523483290046,
          "p25": 26.867685119999805,
          "p75": 57.83864623098634
        },
        "above_30_seconds": 80,
        "bins": [
          0,
          0,
          11,
          29,
          80
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 1406.5666666666666,
        "median": 1069.5,
        "p95": 2710,
        "maximum": 8459
      },
      "input_tokens": {
        "count": 120,
        "mean": 123.20833333333333,
        "median": 115.0,
        "p95": 181,
        "maximum": 295
      },
      "thinking_tokens": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "server_tokens_per_second": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "first_token_seconds": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "no_final": 0
    },
    {
      "configuration": "granite4.2-30b/off",
      "condition": "supplied_passage",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 95.24166666666666,
        "fully_correct": 62,
        "profile_assessed": 115,
        "mean_fully_met_claim_coverage": 0.9344444444444444,
        "answers_with_material_errors": 52,
        "material_error_assessed": 115,
        "pending_reviews": 8,
        "operationally_flagged_answers": 1,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 1,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 825.6416666666667,
        "mean_client_seconds": 33.9371785978173
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 33.9371785978173,
        "median_seconds": 27.0714332810021,
        "p90_seconds": 56.9543606749794,
        "p95_seconds": 72.32749069301644,
        "p99_seconds": 120.64004565699724,
        "max_seconds": 157.4918767880008,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 27.0714332810021,
        "final_returned_p95_seconds": 72.32749069301644,
        "first_token_timing_cases": 0,
        "first_token_p95_seconds": null,
        "observed_above_seconds": {
          "120": {
            "count": 2,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 20.237421446974622,
        "p75_seconds": 39.095746990991756,
        "final_answer_completion": {
          "count": 120,
          "mean": 33.9371785978173,
          "median": 27.0714332810021,
          "p95": 72.32749069301644,
          "maximum": 157.4918767880008,
          "p25": 20.237421446974622,
          "p75": 39.095746990991756
        },
        "above_30_seconds": 51,
        "bins": [
          0,
          2,
          26,
          41,
          51
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 825.6416666666667,
        "median": 682.5,
        "p95": 1731,
        "maximum": 3095
      },
      "input_tokens": {
        "count": 120,
        "mean": 358.39166666666665,
        "median": 346.5,
        "p95": 578,
        "maximum": 1005
      },
      "thinking_tokens": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "server_tokens_per_second": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "first_token_seconds": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "no_final": 0
    },
    {
      "configuration": "lfm2.5-24b-a2b/native",
      "condition": "no_source",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 41.925,
        "fully_correct": 7,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.3566666666666667,
        "answers_with_material_errors": 113,
        "material_error_assessed": 120,
        "pending_reviews": 2,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 5959.275,
        "mean_client_seconds": 28.164994620691385
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 28.164994620691385,
        "median_seconds": 23.837354609480826,
        "p90_seconds": 50.87014818098396,
        "p95_seconds": 65.94923277699854,
        "p99_seconds": 94.2002344270004,
        "max_seconds": 102.41152738500386,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 23.837354609480826,
        "final_returned_p95_seconds": 65.94923277699854,
        "first_token_timing_cases": 120,
        "first_token_p95_seconds": 0.112903,
        "observed_above_seconds": {
          "120": {
            "count": 0,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 15.847320464003133,
        "p75_seconds": 34.345355907978956,
        "final_answer_completion": {
          "count": 120,
          "mean": 28.164994620691385,
          "median": 23.837354609480826,
          "p95": 65.94923277699854,
          "maximum": 102.41152738500386,
          "p25": 15.847320464003133,
          "p75": 34.345355907978956
        },
        "above_30_seconds": 40,
        "bins": [
          0,
          5,
          44,
          31,
          40
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 5959.275,
        "median": 5078.0,
        "p95": 13817,
        "maximum": 20819
      },
      "input_tokens": {
        "count": 120,
        "mean": 125.6,
        "median": 117.0,
        "p95": 182,
        "maximum": 302
      },
      "thinking_tokens": {
        "count": 120,
        "mean": 5615.525,
        "median": 4799.5,
        "p95": 13256,
        "maximum": 20484
      },
      "server_tokens_per_second": {
        "count": 120,
        "mean": 214.9071057102925,
        "median": 215.82402574523343,
        "p95": 220.9844143580983,
        "maximum": 223.11112527689684
      },
      "first_token_seconds": {
        "count": 120,
        "mean": 0.09672036666666667,
        "median": 0.09411,
        "p95": 0.112903,
        "maximum": 0.444505
      },
      "no_final": 0
    },
    {
      "configuration": "lfm2.5-24b-a2b/native",
      "condition": "supplied_passage",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 95.075,
        "fully_correct": 73,
        "profile_assessed": 119,
        "mean_fully_met_claim_coverage": 0.9359722222222222,
        "answers_with_material_errors": 45,
        "material_error_assessed": 119,
        "pending_reviews": 1,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 2868.525,
        "mean_client_seconds": 13.50058852727525
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 13.50058852727525,
        "median_seconds": 10.760558850510279,
        "p90_seconds": 22.904596676991787,
        "p95_seconds": 29.947083480015863,
        "p99_seconds": 44.834318874025485,
        "max_seconds": 83.07041241799016,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 10.760558850510279,
        "final_returned_p95_seconds": 29.947083480015863,
        "first_token_timing_cases": 120,
        "first_token_p95_seconds": 0.165519,
        "observed_above_seconds": {
          "120": {
            "count": 0,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 7.902464801009046,
        "p75_seconds": 14.9789040129981,
        "final_answer_completion": {
          "count": 120,
          "mean": 13.50058852727525,
          "median": 10.760558850510279,
          "p95": 29.947083480015863,
          "maximum": 83.07041241799016,
          "p25": 7.902464801009046,
          "p75": 14.9789040129981
        },
        "above_30_seconds": 6,
        "bins": [
          1,
          54,
          48,
          11,
          6
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 2868.525,
        "median": 2335.0,
        "p95": 6361,
        "maximum": 17493
      },
      "input_tokens": {
        "count": 120,
        "mean": 343.3833333333333,
        "median": 333.0,
        "p95": 547,
        "maximum": 962
      },
      "thinking_tokens": {
        "count": 120,
        "mean": 2497.1583333333333,
        "median": 1901.0,
        "p95": 5920,
        "maximum": 16991
      },
      "server_tokens_per_second": {
        "count": 120,
        "mean": 216.69947031216475,
        "median": 217.2062374712121,
        "p95": 221.5937230035335,
        "maximum": 222.83188274353992
      },
      "first_token_seconds": {
        "count": 120,
        "mean": 0.12212041666666666,
        "median": 0.1168755,
        "p95": 0.165519,
        "maximum": 0.255921
      },
      "no_final": 0
    },
    {
      "configuration": "qwen3.8-27b/off",
      "condition": "no_source",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 53.666666666666664,
        "fully_correct": 6,
        "profile_assessed": 119,
        "mean_fully_met_claim_coverage": 0.47333333333333333,
        "answers_with_material_errors": 112,
        "material_error_assessed": 119,
        "pending_reviews": 14,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 836.3,
        "mean_client_seconds": 16.345276489981188
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 16.345276489981188,
        "median_seconds": 14.209506271989085,
        "p90_seconds": 22.84523086398258,
        "p95_seconds": 28.924308763002045,
        "p99_seconds": 56.505335421999916,
        "max_seconds": 96.13749228202505,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 14.209506271989085,
        "final_returned_p95_seconds": 28.924308763002045,
        "first_token_timing_cases": 0,
        "first_token_p95_seconds": null,
        "observed_above_seconds": {
          "120": {
            "count": 0,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 11.339582551998319,
        "p75_seconds": 18.453588450996904,
        "final_answer_completion": {
          "count": 120,
          "mean": 16.345276489981188,
          "median": 14.209506271989085,
          "p95": 28.924308763002045,
          "maximum": 96.13749228202505,
          "p25": 11.339582551998319,
          "p75": 18.453588450996904
        },
        "above_30_seconds": 6,
        "bins": [
          0,
          21,
          74,
          19,
          6
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 836.3,
        "median": 730.0,
        "p95": 1531,
        "maximum": 5050
      },
      "input_tokens": {
        "count": 120,
        "mean": 124.375,
        "median": 116.0,
        "p95": 182,
        "maximum": 297
      },
      "thinking_tokens": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "server_tokens_per_second": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "first_token_seconds": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "no_final": 0
    },
    {
      "configuration": "qwen3.8-27b/off",
      "condition": "supplied_passage",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 95.15833333333333,
        "fully_correct": 90,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.9366666666666666,
        "answers_with_material_errors": 28,
        "material_error_assessed": 120,
        "pending_reviews": 1,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 456.84166666666664,
        "mean_client_seconds": 8.58779860207481
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 8.58779860207481,
        "median_seconds": 7.705773909488926,
        "p90_seconds": 13.217989357013721,
        "p95_seconds": 14.14093909598887,
        "p99_seconds": 28.765928587992676,
        "max_seconds": 33.19555814098567,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 7.705773909488926,
        "final_returned_p95_seconds": 14.14093909598887,
        "first_token_timing_cases": 0,
        "first_token_p95_seconds": null,
        "observed_above_seconds": {
          "120": {
            "count": 0,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 5.67505301698111,
        "p75_seconds": 10.347955735982396,
        "final_answer_completion": {
          "count": 120,
          "mean": 8.58779860207481,
          "median": 7.705773909488926,
          "p95": 14.14093909598887,
          "maximum": 33.19555814098567,
          "p25": 5.67505301698111,
          "p75": 10.347955735982396
        },
        "above_30_seconds": 1,
        "bins": [
          20,
          67,
          30,
          2,
          1
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 456.84166666666664,
        "median": 406.5,
        "p95": 817,
        "maximum": 1990
      },
      "input_tokens": {
        "count": 120,
        "mean": 349.78333333333336,
        "median": 337.5,
        "p95": 556,
        "maximum": 964
      },
      "thinking_tokens": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "server_tokens_per_second": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "first_token_seconds": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "no_final": 0
    },
    {
      "configuration": "qwen3.8-27b/on",
      "condition": "no_source",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 53.875,
        "fully_correct": 12,
        "profile_assessed": 119,
        "mean_fully_met_claim_coverage": 0.4751388888888889,
        "answers_with_material_errors": 106,
        "material_error_assessed": 119,
        "pending_reviews": 11,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 1735.025,
        "mean_client_seconds": 31.433135336017827
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 31.433135336017827,
        "median_seconds": 27.739497161994223,
        "p90_seconds": 49.86874774500029,
        "p95_seconds": 68.27020600097603,
        "p99_seconds": 104.80797892299597,
        "max_seconds": 105.94480921301874,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 27.739497161994223,
        "final_returned_p95_seconds": 68.27020600097603,
        "first_token_timing_cases": 0,
        "first_token_p95_seconds": null,
        "observed_above_seconds": {
          "120": {
            "count": 0,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 20.84977665601764,
        "p75_seconds": 37.69980688201031,
        "final_answer_completion": {
          "count": 120,
          "mean": 31.433135336017827,
          "median": 27.739497161994223,
          "p95": 68.27020600097603,
          "maximum": 105.94480921301874,
          "p25": 20.84977665601764,
          "p75": 37.69980688201031
        },
        "above_30_seconds": 52,
        "bins": [
          0,
          1,
          27,
          40,
          52
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 1735.025,
        "median": 1513.0,
        "p95": 4129,
        "maximum": 6301
      },
      "input_tokens": {
        "count": 120,
        "mean": 148.375,
        "median": 140.0,
        "p95": 206,
        "maximum": 321
      },
      "thinking_tokens": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "server_tokens_per_second": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "first_token_seconds": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "no_final": 0
    },
    {
      "configuration": "qwen3.8-27b/on",
      "condition": "supplied_passage",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 98.41666666666667,
        "fully_correct": 102,
        "profile_assessed": 119,
        "mean_fully_met_claim_coverage": 0.98125,
        "answers_with_material_errors": 16,
        "material_error_assessed": 119,
        "pending_reviews": 2,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 709.6416666666667,
        "mean_client_seconds": 12.722110486893508
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 12.722110486893508,
        "median_seconds": 11.276302856014809,
        "p90_seconds": 18.20637133301352,
        "p95_seconds": 22.271306298003765,
        "p99_seconds": 32.96603372899699,
        "max_seconds": 44.50844476700877,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 11.276302856014809,
        "final_returned_p95_seconds": 22.271306298003765,
        "first_token_timing_cases": 0,
        "first_token_p95_seconds": null,
        "observed_above_seconds": {
          "120": {
            "count": 0,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 9.815797462011687,
        "p75_seconds": 14.364686550979968,
        "final_answer_completion": {
          "count": 120,
          "mean": 12.722110486893508,
          "median": 11.276302856014809,
          "p95": 22.271306298003765,
          "maximum": 44.50844476700877,
          "p25": 9.815797462011687,
          "p75": 14.364686550979968
        },
        "above_30_seconds": 2,
        "bins": [
          0,
          33,
          77,
          8,
          2
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 709.6416666666667,
        "median": 632.0,
        "p95": 1247,
        "maximum": 2659
      },
      "input_tokens": {
        "count": 120,
        "mean": 373.78333333333336,
        "median": 361.5,
        "p95": 580,
        "maximum": 988
      },
      "thinking_tokens": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "server_tokens_per_second": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "first_token_seconds": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "no_final": 0
    },
    {
      "configuration": "gemma4-26b-a4b/off",
      "condition": "no_source",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 48.71666666666667,
        "fully_correct": 6,
        "profile_assessed": 118,
        "mean_fully_met_claim_coverage": 0.4215277777777778,
        "answers_with_material_errors": 111,
        "material_error_assessed": 118,
        "pending_reviews": 5,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 389.2083333333333,
        "mean_client_seconds": 3.2954716571415097
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 3.2954716571415097,
        "median_seconds": 3.1383249454956967,
        "p90_seconds": 5.036440906987991,
        "p95_seconds": 5.9794541420124006,
        "p99_seconds": 7.333905334002338,
        "max_seconds": 7.601430270005949,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 3.1383249454956967,
        "final_returned_p95_seconds": 5.9794541420124006,
        "first_token_timing_cases": 120,
        "first_token_p95_seconds": 0.24919300000000003,
        "observed_above_seconds": {
          "120": {
            "count": 0,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 2.3658776989905164,
        "p75_seconds": 3.8282598789955955,
        "final_answer_completion": {
          "count": 120,
          "mean": 3.2954716571415097,
          "median": 3.1383249454956967,
          "p95": 5.9794541420124006,
          "maximum": 7.601430270005949,
          "p25": 2.3658776989905164,
          "p75": 3.8282598789955955
        },
        "above_30_seconds": 0,
        "bins": [
          107,
          13,
          0,
          0,
          0
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 389.2083333333333,
        "median": 371.5,
        "p95": 634,
        "maximum": 948
      },
      "input_tokens": {
        "count": 120,
        "mean": 126.33333333333333,
        "median": 118.0,
        "p95": 185,
        "maximum": 296
      },
      "thinking_tokens": {
        "count": 120,
        "mean": 0,
        "median": 0.0,
        "p95": 0,
        "maximum": 0
      },
      "server_tokens_per_second": {
        "count": 120,
        "mean": 133.8167892125082,
        "median": 135.06990219194302,
        "p95": 136.08178266424682,
        "maximum": 136.65570255277075
      },
      "first_token_seconds": {
        "count": 120,
        "mean": 0.11510863333333334,
        "median": 0.096411,
        "p95": 0.24919300000000003,
        "maximum": 0.43753
      },
      "no_final": 0
    },
    {
      "configuration": "gemma4-26b-a4b/off",
      "condition": "supplied_passage",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 94.59166666666667,
        "fully_correct": 96,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.9215277777777777,
        "answers_with_material_errors": 16,
        "material_error_assessed": 120,
        "pending_reviews": 0,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 255.34166666666667,
        "mean_client_seconds": 2.396756961800323
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 2.396756961800323,
        "median_seconds": 2.1331635170063237,
        "p90_seconds": 3.473940933996346,
        "p95_seconds": 4.555981274024816,
        "p99_seconds": 6.249718392005889,
        "max_seconds": 11.144434504007222,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 2.1331635170063237,
        "final_returned_p95_seconds": 4.555981274024816,
        "first_token_timing_cases": 120,
        "first_token_p95_seconds": 0.322423,
        "observed_above_seconds": {
          "120": {
            "count": 0,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 1.6533574200002477,
        "p75_seconds": 2.731671781017212,
        "final_answer_completion": {
          "count": 120,
          "mean": 2.396756961800323,
          "median": 2.1331635170063237,
          "p95": 4.555981274024816,
          "maximum": 11.144434504007222,
          "p25": 1.6533574200002477,
          "p75": 2.731671781017212
        },
        "above_30_seconds": 0,
        "bins": [
          115,
          4,
          1,
          0,
          0
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 255.34166666666667,
        "median": 235.5,
        "p95": 450,
        "maximum": 610
      },
      "input_tokens": {
        "count": 120,
        "mean": 361.1666666666667,
        "median": 347.5,
        "p95": 579,
        "maximum": 1003
      },
      "thinking_tokens": {
        "count": 120,
        "mean": 0,
        "median": 0.0,
        "p95": 0,
        "maximum": 0
      },
      "server_tokens_per_second": {
        "count": 120,
        "mean": 133.12435755724005,
        "median": 133.70717537918085,
        "p95": 135.77191950868942,
        "maximum": 136.66833140325662
      },
      "first_token_seconds": {
        "count": 120,
        "mean": 0.17529409166666668,
        "median": 0.15581650000000002,
        "p95": 0.322423,
        "maximum": 0.451398
      },
      "no_final": 0
    },
    {
      "configuration": "gemma4-26b-a4b/on",
      "condition": "no_source",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 51.24166666666667,
        "fully_correct": 16,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.4434722222222222,
        "answers_with_material_errors": 102,
        "material_error_assessed": 120,
        "pending_reviews": 2,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 2754.9333333333334,
        "mean_client_seconds": 22.196005501001007
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 22.196005501001007,
        "median_seconds": 17.415960502505186,
        "p90_seconds": 45.56172523699934,
        "p95_seconds": 63.54964979100623,
        "p99_seconds": 71.41336383399903,
        "max_seconds": 72.20774586198968,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 17.415960502505186,
        "final_returned_p95_seconds": 63.54964979100623,
        "first_token_timing_cases": 120,
        "first_token_p95_seconds": 0.126562,
        "observed_above_seconds": {
          "120": {
            "count": 0,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 9.27217273999122,
        "p75_seconds": 30.603828694002004,
        "final_answer_completion": {
          "count": 120,
          "mean": 22.196005501001007,
          "median": 17.415960502505186,
          "p95": 63.54964979100623,
          "maximum": 72.20774586198968,
          "p25": 9.27217273999122,
          "p75": 30.603828694002004
        },
        "above_30_seconds": 32,
        "bins": [
          2,
          32,
          34,
          20,
          32
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 2754.9333333333334,
        "median": 2197.5,
        "p95": 7589,
        "maximum": 8581
      },
      "input_tokens": {
        "count": 120,
        "mean": 124.33333333333333,
        "median": 116.0,
        "p95": 183,
        "maximum": 294
      },
      "thinking_tokens": {
        "count": 120,
        "mean": 2493.0333333333333,
        "median": 1952.0,
        "p95": 7206,
        "maximum": 8357
      },
      "server_tokens_per_second": {
        "count": 120,
        "mean": 128.79115875664795,
        "median": 129.4941155279476,
        "p95": 133.86983360069016,
        "maximum": 134.9780784215145
      },
      "first_token_seconds": {
        "count": 120,
        "mean": 0.09977154166666667,
        "median": 0.094256,
        "p95": 0.126562,
        "maximum": 0.342857
      },
      "no_final": 0
    },
    {
      "configuration": "gemma4-26b-a4b/on",
      "condition": "supplied_passage",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 99.075,
        "fully_correct": 113,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.9770833333333333,
        "answers_with_material_errors": 2,
        "material_error_assessed": 120,
        "pending_reviews": 0,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 892.55,
        "mean_client_seconds": 7.248209042998981
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 7.248209042998981,
        "median_seconds": 6.0843440874887165,
        "p90_seconds": 10.22679254601826,
        "p95_seconds": 15.676330866990611,
        "p99_seconds": 26.857115785998758,
        "max_seconds": 32.93256563300383,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 6.0843440874887165,
        "final_returned_p95_seconds": 15.676330866990611,
        "first_token_timing_cases": 120,
        "first_token_p95_seconds": 0.268415,
        "observed_above_seconds": {
          "120": {
            "count": 0,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 4.822041962994263,
        "p75_seconds": 7.8823426209855825,
        "final_answer_completion": {
          "count": 120,
          "mean": 7.248209042998981,
          "median": 6.0843440874887165,
          "p95": 15.676330866990611,
          "maximum": 32.93256563300383,
          "p25": 4.822041962994263,
          "p75": 7.8823426209855825
        },
        "above_30_seconds": 1,
        "bins": [
          34,
          73,
          8,
          4,
          1
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 892.55,
        "median": 750.0,
        "p95": 1960,
        "maximum": 3981
      },
      "input_tokens": {
        "count": 120,
        "mean": 359.1666666666667,
        "median": 345.5,
        "p95": 577,
        "maximum": 1001
      },
      "thinking_tokens": {
        "count": 120,
        "mean": 683.6583333333333,
        "median": 552.5,
        "p95": 1522,
        "maximum": 3706
      },
      "server_tokens_per_second": {
        "count": 120,
        "mean": 131.52440838515287,
        "median": 131.68561353107324,
        "p95": 134.4730938227515,
        "maximum": 135.85868933817724
      },
      "first_token_seconds": {
        "count": 120,
        "mean": 0.15902423333333332,
        "median": 0.1482705,
        "p95": 0.268415,
        "maximum": 0.346223
      },
      "no_final": 0
    },
    {
      "configuration": "glm4.7-flash/off",
      "condition": "no_source",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 35.86666666666667,
        "fully_correct": 0,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.27125,
        "answers_with_material_errors": 120,
        "material_error_assessed": 120,
        "pending_reviews": 21,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 473.2916666666667,
        "mean_client_seconds": 4.3515448323822055
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 4.3515448323822055,
        "median_seconds": 4.110602660497534,
        "p90_seconds": 6.394233761006035,
        "p95_seconds": 7.721416376007255,
        "p99_seconds": 8.310929497005418,
        "max_seconds": 10.094457867002347,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 4.110602660497534,
        "final_returned_p95_seconds": 7.721416376007255,
        "first_token_timing_cases": 0,
        "first_token_p95_seconds": null,
        "observed_above_seconds": {
          "120": {
            "count": 0,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 3.2855270960135385,
        "p75_seconds": 5.091484102013055,
        "final_answer_completion": {
          "count": 120,
          "mean": 4.3515448323822055,
          "median": 4.110602660497534,
          "p95": 7.721416376007255,
          "maximum": 10.094457867002347,
          "p25": 3.2855270960135385,
          "p75": 5.091484102013055
        },
        "above_30_seconds": 0,
        "bins": [
          86,
          33,
          1,
          0,
          0
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 473.2916666666667,
        "median": 441.5,
        "p95": 856,
        "maximum": 1112
      },
      "input_tokens": {
        "count": 120,
        "mean": 112.725,
        "median": 104.0,
        "p95": 171,
        "maximum": 285
      },
      "thinking_tokens": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "server_tokens_per_second": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "first_token_seconds": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "no_final": 0
    },
    {
      "configuration": "glm4.7-flash/off",
      "condition": "supplied_passage",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 70.80833333333334,
        "fully_correct": 30,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.6493055555555556,
        "answers_with_material_errors": 89,
        "material_error_assessed": 120,
        "pending_reviews": 3,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 356.1666666666667,
        "mean_client_seconds": 3.4040780980246685
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 3.4040780980246685,
        "median_seconds": 3.1147646919998806,
        "p90_seconds": 5.0356341089936905,
        "p95_seconds": 6.068288653012132,
        "p99_seconds": 8.094451915996615,
        "max_seconds": 8.749617649998982,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 3.1147646919998806,
        "final_returned_p95_seconds": 6.068288653012132,
        "first_token_timing_cases": 0,
        "first_token_p95_seconds": null,
        "observed_above_seconds": {
          "120": {
            "count": 0,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 2.410167993977666,
        "p75_seconds": 4.104408179991879,
        "final_answer_completion": {
          "count": 120,
          "mean": 3.4040780980246685,
          "median": 3.1147646919998806,
          "p95": 6.068288653012132,
          "maximum": 8.749617649998982,
          "p25": 2.410167993977666,
          "p75": 4.104408179991879
        },
        "above_30_seconds": 0,
        "bins": [
          107,
          13,
          0,
          0,
          0
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 356.1666666666667,
        "median": 333.5,
        "p95": 643,
        "maximum": 1001
      },
      "input_tokens": {
        "count": 120,
        "mean": 323.84166666666664,
        "median": 313.0,
        "p95": 518,
        "maximum": 917
      },
      "thinking_tokens": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "server_tokens_per_second": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "first_token_seconds": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "no_final": 0
    },
    {
      "configuration": "glm4.7-flash/on",
      "condition": "no_source",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 45.733333333333334,
        "fully_correct": 9,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.38930555555555557,
        "answers_with_material_errors": 111,
        "material_error_assessed": 120,
        "pending_reviews": 8,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 3885.1666666666665,
        "mean_client_seconds": 34.20369698000771
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 34.37950779504099,
        "median_seconds": 28.832616621497436,
        "p90_seconds": 62.33365877100732,
        "p95_seconds": 73.58050612601801,
        "p99_seconds": 107.98939229600364,
        "max_seconds": 142.9352043629915,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 1,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 28.832616621497436,
        "final_returned_p95_seconds": 73.58050612601801,
        "first_token_timing_cases": 0,
        "first_token_p95_seconds": null,
        "observed_above_seconds": {
          "120": {
            "count": 1,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 20.784505132003687,
        "p75_seconds": 40.233277241990436,
        "final_answer_completion": {
          "count": 120,
          "mean": 34.37950779504099,
          "median": 28.832616621497436,
          "p95": 73.58050612601801,
          "maximum": 142.9352043629915,
          "p25": 20.784505132003687,
          "p75": 40.233277241990436
        },
        "above_30_seconds": 56,
        "bins": [
          0,
          0,
          28,
          36,
          56
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 3885.1666666666665,
        "median": 3283.5,
        "p95": 8158,
        "maximum": 15014
      },
      "input_tokens": {
        "count": 120,
        "mean": 112.725,
        "median": 104.0,
        "p95": 171,
        "maximum": 285
      },
      "thinking_tokens": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "server_tokens_per_second": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "first_token_seconds": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "no_final": 0
    },
    {
      "configuration": "glm4.7-flash/on",
      "condition": "supplied_passage",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 90.21666666666667,
        "fully_correct": 82,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.8623611111111111,
        "answers_with_material_errors": 26,
        "material_error_assessed": 120,
        "pending_reviews": 0,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 2061.9166666666665,
        "mean_client_seconds": 17.89113451098286
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 18.117181959207908,
        "median_seconds": 13.765581615487463,
        "p90_seconds": 33.44953015400097,
        "p95_seconds": 44.24636647099396,
        "p99_seconds": 76.92090486900997,
        "max_seconds": 83.05583995699999,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 1,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 13.765581615487463,
        "final_returned_p95_seconds": 44.24636647099396,
        "first_token_timing_cases": 0,
        "first_token_p95_seconds": null,
        "observed_above_seconds": {
          "120": {
            "count": 0,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 11.029520962998504,
        "p75_seconds": 19.1848278180114,
        "final_answer_completion": {
          "count": 120,
          "mean": 18.117181959207908,
          "median": 13.765581615487463,
          "p95": 44.24636647099396,
          "maximum": 83.05583995699999,
          "p25": 11.029520962998504,
          "p75": 19.1848278180114
        },
        "above_30_seconds": 14,
        "bins": [
          0,
          20,
          71,
          15,
          14
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 2061.9166666666665,
        "median": 1604.0,
        "p95": 4544,
        "maximum": 9095
      },
      "input_tokens": {
        "count": 120,
        "mean": 323.84166666666664,
        "median": 313.0,
        "p95": 518,
        "maximum": 917
      },
      "thinking_tokens": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "server_tokens_per_second": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "first_token_seconds": {
        "count": 0,
        "mean": null,
        "median": null,
        "p95": null,
        "maximum": null
      },
      "no_final": 0
    },
    {
      "configuration": "muse-glimmer/low",
      "condition": "no_source",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 60.06666666666667,
        "fully_correct": 11,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.5441666666666667,
        "answers_with_material_errors": 108,
        "material_error_assessed": 120,
        "pending_reviews": 9,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 912.7583333333333,
        "mean_client_seconds": 27.54944677187741
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 27.54944677187741,
        "median_seconds": 27.289599504016223,
        "p90_seconds": 39.13926862101653,
        "p95_seconds": 43.566442504990846,
        "p99_seconds": 46.7977286259993,
        "max_seconds": 51.1879825439828,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 27.289599504016223,
        "final_returned_p95_seconds": 43.566442504990846,
        "first_token_timing_cases": 120,
        "first_token_p95_seconds": 0.265206,
        "observed_above_seconds": {
          "120": {
            "count": 0,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 20.59537050401559,
        "p75_seconds": 31.945420611009467,
        "final_answer_completion": {
          "count": 120,
          "mean": 27.54944677187741,
          "median": 27.289599504016223,
          "p95": 43.566442504990846,
          "maximum": 51.1879825439828,
          "p25": 20.59537050401559,
          "p75": 31.945420611009467
        },
        "above_30_seconds": 42,
        "bins": [
          0,
          0,
          27,
          51,
          42
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 912.7583333333333,
        "median": 904.0,
        "p95": 1446,
        "maximum": 1703
      },
      "input_tokens": {
        "count": 120,
        "mean": 134.85833333333332,
        "median": 126.5,
        "p95": 193,
        "maximum": 308
      },
      "thinking_tokens": {
        "count": 120,
        "mean": 458.40833333333336,
        "median": 418.0,
        "p95": 892,
        "maximum": 1098
      },
      "server_tokens_per_second": {
        "count": 120,
        "mean": 33.415634361740054,
        "median": 33.42487251709115,
        "p95": 33.520467597515044,
        "maximum": 33.573869695141326
      },
      "first_token_seconds": {
        "count": 120,
        "mean": 0.19606734166666667,
        "median": 0.187398,
        "p95": 0.265206,
        "maximum": 0.331065
      },
      "no_final": 0
    },
    {
      "configuration": "muse-glimmer/low",
      "condition": "supplied_passage",
      "records": 120,
      "quality": {
        "planned": 120,
        "answers_returned": 120,
        "operational_failures": 0,
        "graded": 120,
        "exact_claim_scores": 120,
        "mean_claim_score": 98.85833333333333,
        "fully_correct": 96,
        "profile_assessed": 120,
        "mean_fully_met_claim_coverage": 0.9858333333333333,
        "answers_with_material_errors": 24,
        "material_error_assessed": 120,
        "pending_reviews": 3,
        "operationally_flagged_answers": 0,
        "output_ceiling_hits": 0,
        "thinking_marker_breaches": 0,
        "direct_failures": 0,
        "regenerated_answers": 0,
        "mean_output_tokens": 581.0166666666667,
        "mean_client_seconds": 17.81076878534174
      },
      "timing": {
        "recorded_cases": 120,
        "complete_timing_cases": 120,
        "missing_or_partial_timing_cases": 0,
        "mean_seconds": 17.81076878534174,
        "median_seconds": 16.386069722517277,
        "p90_seconds": 24.8872484669846,
        "p95_seconds": 27.455511704989476,
        "p99_seconds": 40.25435937399743,
        "max_seconds": 47.98068379602046,
        "nonempty_finals": 120,
        "no_final_cases": 0,
        "output_ceiling_hits": 0,
        "retried_cases": 0,
        "request_failure_attempts": 0,
        "final_returned_median_seconds": 16.386069722517277,
        "final_returned_p95_seconds": 27.455511704989476,
        "first_token_timing_cases": 120,
        "first_token_p95_seconds": 0.536592,
        "observed_above_seconds": {
          "120": {
            "count": 0,
            "denominator": 120
          },
          "180": {
            "count": 0,
            "denominator": 120
          },
          "300": {
            "count": 0,
            "denominator": 120
          },
          "900": {
            "count": 0,
            "denominator": 120
          },
          "1800": {
            "count": 0,
            "denominator": 120
          }
        },
        "p25_seconds": 13.8378901600081,
        "p75_seconds": 20.646109254012117,
        "final_answer_completion": {
          "count": 120,
          "mean": 17.81076878534174,
          "median": 16.386069722517277,
          "p95": 27.455511704989476,
          "maximum": 47.98068379602046,
          "p25": 13.8378901600081,
          "p75": 20.646109254012117
        },
        "above_30_seconds": 4,
        "bins": [
          0,
          3,
          83,
          30,
          4
        ]
      },
      "output_tokens": {
        "count": 120,
        "mean": 581.0166666666667,
        "median": 534.5,
        "p95": 897,
        "maximum": 1559
      },
      "input_tokens": {
        "count": 120,
        "mean": 345.71666666666664,
        "median": 337.0,
        "p95": 542,
        "maximum": 941
      },
      "thinking_tokens": {
        "count": 120,
        "mean": 215.21666666666667,
        "median": 192.5,
        "p95": 439,
        "maximum": 724
      },
      "server_tokens_per_second": {
        "count": 120,
        "mean": 33.38882234206541,
        "median": 33.40672308043928,
        "p95": 33.50948652411969,
        "maximum": 33.57508405148112
      },
      "first_token_seconds": {
        "count": 120,
        "mean": 0.3745515,
        "median": 0.3600725,
        "p95": 0.536592,
        "maximum": 0.957218
      },
      "no_final": 0
    }
  ],
  "sources": [
    {
      "historical_pilot_sha256": "b3ad837fb975d9033455bd4dadc6d395d0644ab9f25f89064078a80ea7d172c4",
      "suite_id": "tiny",
      "selected_records": 1920,
      "observations_sha256": "1d07de0dd82a2ca520348e2225687220876260b9ceb212f679665159d0d151f4",
      "metrics_sha256": "2489cf08a2d7c23d17464f74dca512b7f352c8188ae59e4266cbccd20357e3a0",
      "runtime_profiles_sha256": "b05668f22eb7b4eae536200890e084dd49cd16ef6055192081d74497aec9f2ae",
      "source_selected_records": 2400
    },
    {
      "historical_pilot_sha256": null,
      "suite_id": "medium",
      "selected_records": 1680,
      "observations_sha256": "1e69f98b64b8893f2ff0770fd648e2f820032beabcd21c3ca3dba780331317ed",
      "metrics_sha256": "a3c50604e69754f261a4c088413ae9f80b5a2946d4b1fec82c0f9ef64524c1f3",
      "runtime_profiles_sha256": "4140e5a517327bed77f0452835bb44ce6739ad657bd1414d406c27d9b938415b",
      "source_selected_records": 1680
    },
    {
      "historical_pilot_sha256": null,
      "suite_id": "large",
      "selected_records": 2400,
      "observations_sha256": "b73550e0657dadb55f31fa62447b1ac56e33dda20403acaeda971c2c5cbb3154",
      "metrics_sha256": "47c7a197d15f3cc849e45cb15ded2c056c3c203c0e7bf1eed9a8ed1648edfa90",
      "runtime_profiles_sha256": "0ba5b1b95b1f5e65090bebc34d6fed90f73f7eace2851f2f9622002d896410e6",
      "source_selected_records": 2400
    }
  ],
  "random_human_audit": "not performed; declined by operator",
  "grader_accuracy": "unmeasured"
}
