{
  "schemaVersion": 1,
  "benchmark": {
    "id": "vals-index",
    "name": "Vals Index · Vals AI",
    "version": "2.1 · private test set",
    "category": "general",
    "question": "Which model does the most economically valuable work?",
    "summary": "A GDP-weighted average of agentic performance across finance, coding, legal, and tax tasks.",
    "source": {
      "name": "Vals AI",
      "url": "https://www.vals.ai/benchmarks/vals_index",
      "methodologyUrl": "https://www.vals.ai/about"
    },
    "measure": "The publisher's composite of agentic evaluations across finance, coding, legal, and tax, weighted by each sector's share of U.S. GDP.",
    "comparisonRule": "Compare within index version 2.1. Versions 1, 2, and 2.1 use different component benchmarks, so their scores are not a series.",
    "limitations": [
      "This is the publisher's composite, not an aicharts ranking. The sector weights are a deliberate simplification of how AI reaches the economy.",
      "The Code Migration component scores a fixed 60-task subset of the published 120-task run, not the full benchmark.",
      "Several components are private evaluations that no independent party can reproduce.",
      "Model names are the identifiers Vals publishes, not aicharts canonical names."
    ],
    "coverage": "charted",
    "tags": [
      "composite",
      "economic impact",
      "agentic",
      "finance",
      "legal",
      "coding",
      "GDP",
      "Vals AI",
      "independent evaluator"
    ]
  },
  "dataset": {
    "benchmarkId": "vals-index",
    "version": "2.1 · private test set",
    "score": {
      "label": "Accuracy",
      "unit": "%",
      "direction": "higher",
      "minimum": 0,
      "maximum": 100
    },
    "source": {
      "name": "Vals AI · Vals Index",
      "url": "https://www.vals.ai/benchmarks/vals_index",
      "retrievedAt": "2026-10-09T16:00:35.747Z",
      "revision": "vals_index v2.1"
    },
    "observedAt": "2026-10-07T00:00:00.000Z",
    "evidenceLabel": "Independent evaluator · private test set",
    "configurationLabel": "Model at the effort Vals selected",
    "comparabilityNote": "One publisher-defined composite. Cost is the average dollar cost of one test across the weighted components, not the cost of a task you would run. Model names are the identifiers Vals publishes, not aicharts canonical names.",
    "costLabel": "Average cost per test (USD)",
    "points": [
      {
        "id": "alibaba-qwen3-8-max",
        "label": "alibaba/qwen3.8-max",
        "model": "alibaba/qwen3.8-max",
        "provider": "Alibaba",
        "harness": null,
        "effort": null,
        "score": 48.27,
        "costUsd": 4.964352,
        "uncertainty": {
          "lower": 46.933,
          "upper": 49.607000000000006,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "alibaba/qwen3.8-max"
          },
          {
            "label": "Standard error",
            "value": "±1.34 points"
          },
          {
            "label": "Mean latency",
            "value": "6,071 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "50.6%"
          },
          {
            "label": "EMB",
            "value": "60.1%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "34.3%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "64.7%"
          },
          {
            "label": "Code Migration",
            "value": "23.2%"
          },
          {
            "label": "Legal Research",
            "value": "47.6%"
          },
          {
            "label": "HLAB",
            "value": "10.4%"
          },
          {
            "label": "Tax Agent",
            "value": "66.0%"
          }
        ]
      },
      {
        "id": "anthropic-claude-fable-5",
        "label": "anthropic/claude-fable-5",
        "model": "anthropic/claude-fable-5",
        "provider": "Anthropic",
        "harness": null,
        "effort": "max",
        "score": 61.388,
        "costUsd": 29.58529,
        "uncertainty": {
          "lower": 60.388999999999996,
          "upper": 62.387,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "anthropic/claude-fable-5"
          },
          {
            "label": "Standard error",
            "value": "±1.00 points"
          },
          {
            "label": "Mean latency",
            "value": "2,544 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "56.3%"
          },
          {
            "label": "EMB",
            "value": "73.7%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "41.4%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "90.4%"
          },
          {
            "label": "Code Migration",
            "value": "54.6%"
          },
          {
            "label": "Legal Research",
            "value": "49.5%"
          },
          {
            "label": "HLAB",
            "value": "11.3%"
          },
          {
            "label": "Tax Agent",
            "value": "69.8%"
          }
        ]
      },
      {
        "id": "anthropic-claude-fable-5-1",
        "label": "anthropic/claude-fable-5-1",
        "model": "anthropic/claude-fable-5-1",
        "provider": "Anthropic",
        "harness": null,
        "effort": "max",
        "score": 65.827,
        "costUsd": 28.711785,
        "uncertainty": {
          "lower": 64.712,
          "upper": 66.942,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "anthropic/claude-fable-5-1"
          },
          {
            "label": "Standard error",
            "value": "±1.11 points"
          },
          {
            "label": "Mean latency",
            "value": "4,612 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "58.9%"
          },
          {
            "label": "EMB",
            "value": "76.7%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "58.1%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "90.3%"
          },
          {
            "label": "Code Migration",
            "value": "60.0%"
          },
          {
            "label": "Legal Research",
            "value": "55.3%"
          },
          {
            "label": "HLAB",
            "value": "6.7%"
          },
          {
            "label": "Tax Agent",
            "value": "77.6%"
          }
        ]
      },
      {
        "id": "anthropic-claude-haiku-5-5",
        "label": "anthropic/claude-haiku-5-5",
        "model": "anthropic/claude-haiku-5-5",
        "provider": "Anthropic",
        "harness": null,
        "effort": "max",
        "score": 54.312,
        "costUsd": 2.991465,
        "uncertainty": {
          "lower": 52.902,
          "upper": 55.721999999999994,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "anthropic/claude-haiku-5-5"
          },
          {
            "label": "Standard error",
            "value": "±1.41 points"
          },
          {
            "label": "Mean latency",
            "value": "2,837 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "54.1%"
          },
          {
            "label": "EMB",
            "value": "59.0%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "35.4%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "90.4%"
          },
          {
            "label": "Code Migration",
            "value": "45.9%"
          },
          {
            "label": "Legal Research",
            "value": "43.3%"
          },
          {
            "label": "HLAB",
            "value": "1.3%"
          },
          {
            "label": "Tax Agent",
            "value": "62.2%"
          }
        ]
      },
      {
        "id": "anthropic-claude-opus-4-8",
        "label": "anthropic/claude-opus-4-8",
        "model": "anthropic/claude-opus-4-8",
        "provider": "Anthropic",
        "harness": null,
        "effort": "max",
        "score": 55.1,
        "costUsd": 13.136067,
        "uncertainty": {
          "lower": 54.091,
          "upper": 56.109,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "anthropic/claude-opus-4-8"
          },
          {
            "label": "Standard error",
            "value": "±1.01 points"
          },
          {
            "label": "Mean latency",
            "value": "2,388 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "53.9%"
          },
          {
            "label": "EMB",
            "value": "69.4%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "23.2%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "82.7%"
          },
          {
            "label": "Code Migration",
            "value": "46.9%"
          },
          {
            "label": "Legal Research",
            "value": "43.8%"
          },
          {
            "label": "HLAB",
            "value": "9.6%"
          },
          {
            "label": "Tax Agent",
            "value": "64.9%"
          }
        ]
      },
      {
        "id": "anthropic-claude-opus-5",
        "label": "anthropic/claude-opus-5",
        "model": "anthropic/claude-opus-5",
        "provider": "Anthropic",
        "harness": null,
        "effort": "max",
        "score": 63.674,
        "costUsd": 19.314752,
        "uncertainty": {
          "lower": 62.717,
          "upper": 64.631,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "anthropic/claude-opus-5"
          },
          {
            "label": "Standard error",
            "value": "±0.96 points"
          },
          {
            "label": "Mean latency",
            "value": "3,462 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "58.6%"
          },
          {
            "label": "EMB",
            "value": "73.6%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "53.5%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "88.4%"
          },
          {
            "label": "Code Migration",
            "value": "56.7%"
          },
          {
            "label": "Legal Research",
            "value": "55.3%"
          },
          {
            "label": "HLAB",
            "value": "6.7%"
          },
          {
            "label": "Tax Agent",
            "value": "75.1%"
          }
        ]
      },
      {
        "id": "anthropic-claude-opus-5-5",
        "label": "anthropic/claude-opus-5-5",
        "model": "anthropic/claude-opus-5-5",
        "provider": "Anthropic",
        "harness": null,
        "effort": "max",
        "score": 66.972,
        "costUsd": 32.143857,
        "uncertainty": {
          "lower": 66.08399999999999,
          "upper": 67.86,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "anthropic/claude-opus-5-5"
          },
          {
            "label": "Standard error",
            "value": "±0.89 points"
          },
          {
            "label": "Mean latency",
            "value": "4,740 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "58.6%"
          },
          {
            "label": "EMB",
            "value": "75.9%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "65.2%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "90.3%"
          },
          {
            "label": "Code Migration",
            "value": "68.9%"
          },
          {
            "label": "Legal Research",
            "value": "50.5%"
          },
          {
            "label": "HLAB",
            "value": "3.8%"
          },
          {
            "label": "Tax Agent",
            "value": "70.5%"
          }
        ]
      },
      {
        "id": "anthropic-claude-sonnet-5",
        "label": "anthropic/claude-sonnet-5",
        "model": "anthropic/claude-sonnet-5",
        "provider": "Anthropic",
        "harness": null,
        "effort": "max",
        "score": 51.775,
        "costUsd": 13.721168,
        "uncertainty": {
          "lower": 50.683,
          "upper": 52.867,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "anthropic/claude-sonnet-5"
          },
          {
            "label": "Standard error",
            "value": "±1.09 points"
          },
          {
            "label": "Mean latency",
            "value": "3,256 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "53.9%"
          },
          {
            "label": "EMB",
            "value": "66.3%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "9.6%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "81.3%"
          },
          {
            "label": "Code Migration",
            "value": "44.1%"
          },
          {
            "label": "Legal Research",
            "value": "41.8%"
          },
          {
            "label": "HLAB",
            "value": "5.0%"
          },
          {
            "label": "Tax Agent",
            "value": "62.3%"
          }
        ]
      },
      {
        "id": "anthropic-claude-sonnet-5-5",
        "label": "anthropic/claude-sonnet-5-5",
        "model": "anthropic/claude-sonnet-5-5",
        "provider": "Anthropic",
        "harness": null,
        "effort": "max",
        "score": 67.037,
        "costUsd": 21.344634,
        "uncertainty": {
          "lower": 66.117,
          "upper": 67.95700000000001,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "anthropic/claude-sonnet-5-5"
          },
          {
            "label": "Standard error",
            "value": "±0.92 points"
          },
          {
            "label": "Mean latency",
            "value": "4,684 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "58.1%"
          },
          {
            "label": "EMB",
            "value": "75.7%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "64.1%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "92.4%"
          },
          {
            "label": "Code Migration",
            "value": "70.1%"
          },
          {
            "label": "Legal Research",
            "value": "48.1%"
          },
          {
            "label": "HLAB",
            "value": "2.9%"
          },
          {
            "label": "Tax Agent",
            "value": "73.4%"
          }
        ]
      },
      {
        "id": "deepseek-deepseek-v4-1-flash",
        "label": "deepseek/deepseek-v4.1-flash",
        "model": "deepseek/deepseek-v4.1-flash",
        "provider": "DeepSeek",
        "harness": null,
        "effort": "high",
        "score": 51.32,
        "costUsd": 0.332286,
        "uncertainty": {
          "lower": 50.19,
          "upper": 52.45,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "deepseek/deepseek-v4.1-flash"
          },
          {
            "label": "Standard error",
            "value": "±1.13 points"
          },
          {
            "label": "Mean latency",
            "value": "1,673 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "53.5%"
          },
          {
            "label": "EMB",
            "value": "57.2%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "19.7%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "84.7%"
          },
          {
            "label": "Code Migration",
            "value": "46.9%"
          },
          {
            "label": "Legal Research",
            "value": "41.3%"
          },
          {
            "label": "HLAB",
            "value": "6.7%"
          },
          {
            "label": "Tax Agent",
            "value": "62.5%"
          }
        ]
      },
      {
        "id": "deepseek-deepseek-v4-flash-0731",
        "label": "deepseek/deepseek-v4-flash-0731",
        "model": "deepseek/deepseek-v4-flash-0731",
        "provider": "DeepSeek",
        "harness": null,
        "effort": "high",
        "score": 47.963,
        "costUsd": 0.847519,
        "uncertainty": {
          "lower": 46.723,
          "upper": 49.203,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "deepseek/deepseek-v4-flash-0731"
          },
          {
            "label": "Standard error",
            "value": "±1.24 points"
          },
          {
            "label": "Mean latency",
            "value": "2,586 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "49.5%"
          },
          {
            "label": "EMB",
            "value": "57.0%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "18.7%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "74.7%"
          },
          {
            "label": "Code Migration",
            "value": "42.3%"
          },
          {
            "label": "Legal Research",
            "value": "30.3%"
          },
          {
            "label": "HLAB",
            "value": "8.3%"
          },
          {
            "label": "Tax Agent",
            "value": "62.6%"
          }
        ]
      },
      {
        "id": "deepseek-deepseek-v4-pro",
        "label": "deepseek/deepseek-v4-pro",
        "model": "deepseek/deepseek-v4-pro",
        "provider": "DeepSeek",
        "harness": null,
        "effort": "max",
        "score": 38.63,
        "costUsd": 1.30566,
        "uncertainty": {
          "lower": 37.473,
          "upper": 39.787000000000006,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "deepseek/deepseek-v4-pro"
          },
          {
            "label": "Standard error",
            "value": "±1.16 points"
          },
          {
            "label": "Mean latency",
            "value": "1,929 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "44.1%"
          },
          {
            "label": "EMB",
            "value": "51.6%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "11.1%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "49.9%"
          },
          {
            "label": "Code Migration",
            "value": "26.2%"
          },
          {
            "label": "Legal Research",
            "value": "23.1%"
          },
          {
            "label": "HLAB",
            "value": "3.8%"
          },
          {
            "label": "Tax Agent",
            "value": "58.5%"
          }
        ]
      },
      {
        "id": "deepseek-deepseek-v4-pro-0813",
        "label": "deepseek/deepseek-v4-pro-0813",
        "model": "deepseek/deepseek-v4-pro-0813",
        "provider": "DeepSeek",
        "harness": null,
        "effort": "max",
        "score": 47.628,
        "costUsd": 3.352436,
        "uncertainty": {
          "lower": 46.512,
          "upper": 48.744,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "deepseek/deepseek-v4-pro-0813"
          },
          {
            "label": "Standard error",
            "value": "±1.12 points"
          },
          {
            "label": "Mean latency",
            "value": "3,845 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "50.4%"
          },
          {
            "label": "EMB",
            "value": "52.8%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "14.1%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "82.3%"
          },
          {
            "label": "Code Migration",
            "value": "41.5%"
          },
          {
            "label": "Legal Research",
            "value": "40.9%"
          },
          {
            "label": "HLAB",
            "value": "7.5%"
          },
          {
            "label": "Tax Agent",
            "value": "58.7%"
          }
        ]
      },
      {
        "id": "fireworks-ember-1",
        "label": "fireworks/ember-1",
        "model": "fireworks/ember-1",
        "provider": "Fireworks AI",
        "harness": null,
        "effort": null,
        "score": 50.811,
        "costUsd": 9.577707,
        "uncertainty": {
          "lower": 49.534,
          "upper": 52.088,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "fireworks/ember-1"
          },
          {
            "label": "Standard error",
            "value": "±1.28 points"
          },
          {
            "label": "Mean latency",
            "value": "4,261 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "51.9%"
          },
          {
            "label": "EMB",
            "value": "63.4%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "19.7%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "83.9%"
          },
          {
            "label": "Code Migration",
            "value": "30.5%"
          },
          {
            "label": "Legal Research",
            "value": "44.7%"
          },
          {
            "label": "HLAB",
            "value": "9.2%"
          },
          {
            "label": "Tax Agent",
            "value": "68.0%"
          }
        ]
      },
      {
        "id": "google-gemini-3-1-pro-preview",
        "label": "google/gemini-3.1-pro-preview",
        "model": "google/gemini-3.1-pro-preview",
        "provider": "Google",
        "harness": null,
        "effort": "high",
        "score": 33.442,
        "costUsd": 2.180322,
        "uncertainty": {
          "lower": 32.309,
          "upper": 34.575,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "google/gemini-3.1-pro-preview"
          },
          {
            "label": "Standard error",
            "value": "±1.13 points"
          },
          {
            "label": "Mean latency",
            "value": "592 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "43.0%"
          },
          {
            "label": "EMB",
            "value": "52.6%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "2.5%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "32.0%"
          },
          {
            "label": "Code Migration",
            "value": "17.9%"
          },
          {
            "label": "Legal Research",
            "value": "20.7%"
          },
          {
            "label": "HLAB",
            "value": "0.0%"
          },
          {
            "label": "Tax Agent",
            "value": "37.8%"
          }
        ]
      },
      {
        "id": "google-gemini-3-5-flash",
        "label": "google/gemini-3.5-flash",
        "model": "google/gemini-3.5-flash",
        "provider": "Google",
        "harness": null,
        "effort": "high",
        "score": 44.786,
        "costUsd": 3.297245,
        "uncertainty": {
          "lower": 43.688,
          "upper": 45.884,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "google/gemini-3.5-flash"
          },
          {
            "label": "Standard error",
            "value": "±1.10 points"
          },
          {
            "label": "Mean latency",
            "value": "1,097 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "57.9%"
          },
          {
            "label": "EMB",
            "value": "63.5%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "6.1%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "48.7%"
          },
          {
            "label": "Code Migration",
            "value": "27.1%"
          },
          {
            "label": "Legal Research",
            "value": "30.8%"
          },
          {
            "label": "HLAB",
            "value": "2.5%"
          },
          {
            "label": "Tax Agent",
            "value": "53.5%"
          }
        ]
      },
      {
        "id": "google-gemini-3-7-flash",
        "label": "google/gemini-3.7-flash",
        "model": "google/gemini-3.7-flash",
        "provider": "Google",
        "harness": null,
        "effort": "high",
        "score": 51.273,
        "costUsd": 4.708904,
        "uncertainty": {
          "lower": 50.176,
          "upper": 52.370000000000005,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "google/gemini-3.7-flash"
          },
          {
            "label": "Standard error",
            "value": "±1.10 points"
          },
          {
            "label": "Mean latency",
            "value": "2,847 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "59.0%"
          },
          {
            "label": "EMB",
            "value": "71.3%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "12.1%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "70.4%"
          },
          {
            "label": "Code Migration",
            "value": "29.0%"
          },
          {
            "label": "Legal Research",
            "value": "34.6%"
          },
          {
            "label": "HLAB",
            "value": "8.8%"
          },
          {
            "label": "Tax Agent",
            "value": "57.7%"
          }
        ]
      },
      {
        "id": "google-gemini-3-8-flash",
        "label": "google/gemini-3.8-flash",
        "model": "google/gemini-3.8-flash",
        "provider": "Google",
        "harness": null,
        "effort": "high",
        "score": 54.825,
        "costUsd": 5.728779,
        "uncertainty": {
          "lower": 53.797000000000004,
          "upper": 55.853,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "google/gemini-3.8-flash"
          },
          {
            "label": "Standard error",
            "value": "±1.03 points"
          },
          {
            "label": "Mean latency",
            "value": "3,117 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "61.4%"
          },
          {
            "label": "EMB",
            "value": "72.2%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "19.2%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "78.7%"
          },
          {
            "label": "Code Migration",
            "value": "31.6%"
          },
          {
            "label": "Legal Research",
            "value": "38.9%"
          },
          {
            "label": "HLAB",
            "value": "10.0%"
          },
          {
            "label": "Tax Agent",
            "value": "66.8%"
          }
        ]
      },
      {
        "id": "google-gemini-4-argon",
        "label": "google/gemini-4-argon",
        "model": "google/gemini-4-argon",
        "provider": "Google",
        "harness": null,
        "effort": "high",
        "score": 68.896,
        "costUsd": 15.681703,
        "uncertainty": {
          "lower": 67.922,
          "upper": 69.87,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "google/gemini-4-argon"
          },
          {
            "label": "Standard error",
            "value": "±0.97 points"
          },
          {
            "label": "Mean latency",
            "value": "2,793 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "65.4%"
          },
          {
            "label": "EMB",
            "value": "75.2%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "57.6%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "91.9%"
          },
          {
            "label": "Code Migration",
            "value": "69.5%"
          },
          {
            "label": "Legal Research",
            "value": "54.8%"
          },
          {
            "label": "HLAB",
            "value": "19.6%"
          },
          {
            "label": "Tax Agent",
            "value": "76.2%"
          }
        ]
      },
      {
        "id": "grok-grok-4-5",
        "label": "grok/grok-4.5",
        "model": "grok/grok-4.5",
        "provider": "SpaceXAI",
        "harness": null,
        "effort": "high",
        "score": 44.723,
        "costUsd": 2.458898,
        "uncertainty": {
          "lower": 43.552,
          "upper": 45.894,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "grok/grok-4.5"
          },
          {
            "label": "Standard error",
            "value": "±1.17 points"
          },
          {
            "label": "Mean latency",
            "value": "944 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "48.3%"
          },
          {
            "label": "EMB",
            "value": "52.9%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "8.6%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "69.0%"
          },
          {
            "label": "Code Migration",
            "value": "38.4%"
          },
          {
            "label": "Legal Research",
            "value": "38.0%"
          },
          {
            "label": "HLAB",
            "value": "12.9%"
          },
          {
            "label": "Tax Agent",
            "value": "64.3%"
          }
        ]
      },
      {
        "id": "grok-grok-4-6",
        "label": "grok/grok-4.6",
        "model": "grok/grok-4.6",
        "provider": "SpaceXAI",
        "harness": null,
        "effort": "high",
        "score": 52.092,
        "costUsd": 4.49288,
        "uncertainty": {
          "lower": 50.939,
          "upper": 53.245,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "grok/grok-4.6"
          },
          {
            "label": "Standard error",
            "value": "±1.15 points"
          },
          {
            "label": "Mean latency",
            "value": "2,221 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "53.7%"
          },
          {
            "label": "EMB",
            "value": "62.7%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "17.2%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "76.2%"
          },
          {
            "label": "Code Migration",
            "value": "44.6%"
          },
          {
            "label": "Legal Research",
            "value": "48.1%"
          },
          {
            "label": "HLAB",
            "value": "15.8%"
          },
          {
            "label": "Tax Agent",
            "value": "70.8%"
          }
        ]
      },
      {
        "id": "grok-grok-4-7",
        "label": "grok/grok-4.7",
        "model": "grok/grok-4.7",
        "provider": "SpaceXAI",
        "harness": null,
        "effort": "xhigh",
        "score": 54.947,
        "costUsd": 12.118089,
        "uncertainty": {
          "lower": 53.875,
          "upper": 56.019000000000005,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "grok/grok-4.7"
          },
          {
            "label": "Standard error",
            "value": "±1.07 points"
          },
          {
            "label": "Mean latency",
            "value": "2,131 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "52.3%"
          },
          {
            "label": "EMB",
            "value": "67.0%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "28.8%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "86.2%"
          },
          {
            "label": "Code Migration",
            "value": "43.2%"
          },
          {
            "label": "Legal Research",
            "value": "47.1%"
          },
          {
            "label": "HLAB",
            "value": "12.5%"
          },
          {
            "label": "Tax Agent",
            "value": "65.6%"
          }
        ]
      },
      {
        "id": "inception-mercury-2-5",
        "label": "inception/mercury-2.5",
        "model": "inception/mercury-2.5",
        "provider": "Inception",
        "harness": null,
        "effort": "high",
        "score": 8.855,
        "costUsd": 0.285616,
        "uncertainty": {
          "lower": 8.409,
          "upper": 9.301,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "inception/mercury-2.5"
          },
          {
            "label": "Standard error",
            "value": "±0.45 points"
          },
          {
            "label": "Mean latency",
            "value": "342 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "18.9%"
          },
          {
            "label": "EMB",
            "value": "8.8%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "0.0%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "3.7%"
          },
          {
            "label": "Code Migration",
            "value": "4.9%"
          },
          {
            "label": "Legal Research",
            "value": "4.3%"
          },
          {
            "label": "HLAB",
            "value": "0.0%"
          },
          {
            "label": "Tax Agent",
            "value": "12.4%"
          }
        ]
      },
      {
        "id": "kimi-kimi-k3",
        "label": "kimi/kimi-k3",
        "model": "kimi/kimi-k3",
        "provider": "Moonshot AI",
        "harness": null,
        "effort": "max",
        "score": 50.297,
        "costUsd": 6.382152,
        "uncertainty": {
          "lower": 49.305,
          "upper": 51.288999999999994,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "kimi/kimi-k3"
          },
          {
            "label": "Standard error",
            "value": "±0.99 points"
          },
          {
            "label": "Mean latency",
            "value": "4,103 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "53.1%"
          },
          {
            "label": "EMB",
            "value": "66.7%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "17.2%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "85.0%"
          },
          {
            "label": "Code Migration",
            "value": "16.0%"
          },
          {
            "label": "Legal Research",
            "value": "46.2%"
          },
          {
            "label": "HLAB",
            "value": "12.9%"
          },
          {
            "label": "Tax Agent",
            "value": "68.7%"
          }
        ]
      },
      {
        "id": "meta-muse-spark-1-2",
        "label": "meta/muse_spark_1_2",
        "model": "meta/muse_spark_1_2",
        "provider": "Meta",
        "harness": null,
        "effort": "xhigh",
        "score": 49.287,
        "costUsd": 1.942052,
        "uncertainty": {
          "lower": 48.191,
          "upper": 50.382999999999996,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "meta/muse_spark_1_2"
          },
          {
            "label": "Standard error",
            "value": "±1.10 points"
          },
          {
            "label": "Mean latency",
            "value": "1,376 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "60.6%"
          },
          {
            "label": "EMB",
            "value": "57.0%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "6.1%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "79.1%"
          },
          {
            "label": "Code Migration",
            "value": "29.4%"
          },
          {
            "label": "Legal Research",
            "value": "43.8%"
          },
          {
            "label": "HLAB",
            "value": "25.4%"
          },
          {
            "label": "Tax Agent",
            "value": "56.9%"
          }
        ]
      },
      {
        "id": "meta-muse-spark-1-3",
        "label": "meta/muse_spark_1_3",
        "model": "meta/muse_spark_1_3",
        "provider": "Meta",
        "harness": null,
        "effort": "xhigh",
        "score": 53.197,
        "costUsd": 3.37642,
        "uncertainty": {
          "lower": 52.131,
          "upper": 54.263000000000005,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "meta/muse_spark_1_3"
          },
          {
            "label": "Standard error",
            "value": "±1.07 points"
          },
          {
            "label": "Mean latency",
            "value": "1,966 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "58.9%"
          },
          {
            "label": "EMB",
            "value": "62.7%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "10.6%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "82.9%"
          },
          {
            "label": "Code Migration",
            "value": "42.2%"
          },
          {
            "label": "Legal Research",
            "value": "40.9%"
          },
          {
            "label": "HLAB",
            "value": "22.9%"
          },
          {
            "label": "Tax Agent",
            "value": "71.9%"
          }
        ]
      },
      {
        "id": "meta-muse-spark-1-3-max",
        "label": "meta/muse_spark_1_3_max",
        "model": "meta/muse_spark_1_3_max",
        "provider": "Meta",
        "harness": null,
        "effort": "max",
        "score": 58.162,
        "costUsd": 3.787017,
        "uncertainty": {
          "lower": 56.974,
          "upper": 59.35,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "meta/muse_spark_1_3_max"
          },
          {
            "label": "Standard error",
            "value": "±1.19 points"
          },
          {
            "label": "Mean latency",
            "value": "1,414 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "60.0%"
          },
          {
            "label": "EMB",
            "value": "67.4%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "24.7%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "85.9%"
          },
          {
            "label": "Code Migration",
            "value": "48.3%"
          },
          {
            "label": "Legal Research",
            "value": "55.3%"
          },
          {
            "label": "HLAB",
            "value": "23.8%"
          },
          {
            "label": "Tax Agent",
            "value": "72.4%"
          }
        ]
      },
      {
        "id": "minimax-minimax-m3",
        "label": "minimax/MiniMax-M3",
        "model": "minimax/MiniMax-M3",
        "provider": "MiniMax",
        "harness": null,
        "effort": null,
        "score": 36.535,
        "costUsd": 3.082399,
        "uncertainty": {
          "lower": 35.367999999999995,
          "upper": 37.702,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "minimax/MiniMax-M3"
          },
          {
            "label": "Standard error",
            "value": "±1.17 points"
          },
          {
            "label": "Mean latency",
            "value": "2,335 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "48.3%"
          },
          {
            "label": "EMB",
            "value": "47.8%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "1.0%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "47.6%"
          },
          {
            "label": "Code Migration",
            "value": "20.9%"
          },
          {
            "label": "Legal Research",
            "value": "29.8%"
          },
          {
            "label": "HLAB",
            "value": "4.2%"
          },
          {
            "label": "Tax Agent",
            "value": "49.7%"
          }
        ]
      },
      {
        "id": "mistralai-mistral-large-4",
        "label": "mistralai/mistral-large-4",
        "model": "mistralai/mistral-large-4",
        "provider": "Mistral AI",
        "harness": null,
        "effort": "high",
        "score": 48.045,
        "costUsd": 13.776396,
        "uncertainty": {
          "lower": 46.933,
          "upper": 49.157000000000004,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "mistralai/mistral-large-4"
          },
          {
            "label": "Standard error",
            "value": "±1.11 points"
          },
          {
            "label": "Mean latency",
            "value": "6,350 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "54.7%"
          },
          {
            "label": "EMB",
            "value": "56.2%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "22.7%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "78.4%"
          },
          {
            "label": "Code Migration",
            "value": "22.8%"
          },
          {
            "label": "Legal Research",
            "value": "31.7%"
          },
          {
            "label": "HLAB",
            "value": "15.8%"
          },
          {
            "label": "Tax Agent",
            "value": "63.3%"
          }
        ]
      },
      {
        "id": "openai-gpt-5-4-mini-2026-03-17",
        "label": "openai/gpt-5.4-mini-2026-03-17",
        "model": "openai/gpt-5.4-mini-2026-03-17",
        "provider": "OpenAI",
        "harness": null,
        "effort": "xhigh",
        "score": 33.169,
        "costUsd": 1.45173,
        "uncertainty": {
          "lower": 31.886999999999997,
          "upper": 34.45099999999999,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "openai/gpt-5.4-mini-2026-03-17"
          },
          {
            "label": "Standard error",
            "value": "±1.28 points"
          },
          {
            "label": "Mean latency",
            "value": "1,789 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "45.4%"
          },
          {
            "label": "EMB",
            "value": "45.4%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "2.5%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "48.0%"
          },
          {
            "label": "Code Migration",
            "value": "13.0%"
          },
          {
            "label": "Legal Research",
            "value": "12.5%"
          },
          {
            "label": "HLAB",
            "value": "0.0%"
          },
          {
            "label": "Tax Agent",
            "value": "36.6%"
          }
        ]
      },
      {
        "id": "openai-gpt-5-6-luna",
        "label": "openai/gpt-5.6-luna",
        "model": "openai/gpt-5.6-luna",
        "provider": "OpenAI",
        "harness": null,
        "effort": "max",
        "score": 51.687,
        "costUsd": 0.824394,
        "uncertainty": {
          "lower": 50.629999999999995,
          "upper": 52.744,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "openai/gpt-5.6-luna"
          },
          {
            "label": "Standard error",
            "value": "±1.06 points"
          },
          {
            "label": "Mean latency",
            "value": "1,767 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "55.0%"
          },
          {
            "label": "EMB",
            "value": "67.1%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "11.6%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "77.1%"
          },
          {
            "label": "Code Migration",
            "value": "44.8%"
          },
          {
            "label": "Legal Research",
            "value": "36.5%"
          },
          {
            "label": "HLAB",
            "value": "1.3%"
          },
          {
            "label": "Tax Agent",
            "value": "60.8%"
          }
        ]
      },
      {
        "id": "openai-gpt-5-6-sol",
        "label": "openai/gpt-5.6-sol",
        "model": "openai/gpt-5.6-sol",
        "provider": "OpenAI",
        "harness": null,
        "effort": "max",
        "score": 58.005,
        "costUsd": 14.239384,
        "uncertainty": {
          "lower": 56.988,
          "upper": 59.022000000000006,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "openai/gpt-5.6-sol"
          },
          {
            "label": "Standard error",
            "value": "±1.02 points"
          },
          {
            "label": "Mean latency",
            "value": "2,243 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "53.8%"
          },
          {
            "label": "EMB",
            "value": "72.3%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "37.9%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "80.5%"
          },
          {
            "label": "Code Migration",
            "value": "52.4%"
          },
          {
            "label": "Legal Research",
            "value": "48.1%"
          },
          {
            "label": "HLAB",
            "value": "2.5%"
          },
          {
            "label": "Tax Agent",
            "value": "68.0%"
          }
        ]
      },
      {
        "id": "openai-gpt-5-6-terra",
        "label": "openai/gpt-5.6-terra",
        "model": "openai/gpt-5.6-terra",
        "provider": "OpenAI",
        "harness": null,
        "effort": "max",
        "score": 53.085,
        "costUsd": 5.776241,
        "uncertainty": {
          "lower": 51.792,
          "upper": 54.378,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "openai/gpt-5.6-terra"
          },
          {
            "label": "Standard error",
            "value": "±1.29 points"
          },
          {
            "label": "Mean latency",
            "value": "2,139 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "54.4%"
          },
          {
            "label": "EMB",
            "value": "66.2%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "22.7%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "74.6%"
          },
          {
            "label": "Code Migration",
            "value": "48.2%"
          },
          {
            "label": "Legal Research",
            "value": "41.3%"
          },
          {
            "label": "HLAB",
            "value": "0.8%"
          },
          {
            "label": "Tax Agent",
            "value": "65.2%"
          }
        ]
      },
      {
        "id": "openai-gpt-6-1-sol",
        "label": "openai/gpt-6.1-sol",
        "model": "openai/gpt-6.1-sol",
        "provider": "OpenAI",
        "harness": null,
        "effort": "max",
        "score": 61.154,
        "costUsd": 3.236975,
        "uncertainty": {
          "lower": 60.143,
          "upper": 62.165000000000006,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "openai/gpt-6.1-sol"
          },
          {
            "label": "Standard error",
            "value": "±1.01 points"
          },
          {
            "label": "Mean latency",
            "value": "2,590 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "52.0%"
          },
          {
            "label": "EMB",
            "value": "70.8%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "55.1%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "88.9%"
          },
          {
            "label": "Code Migration",
            "value": "63.3%"
          },
          {
            "label": "Legal Research",
            "value": "38.5%"
          },
          {
            "label": "HLAB",
            "value": "5.4%"
          },
          {
            "label": "Tax Agent",
            "value": "62.3%"
          }
        ]
      },
      {
        "id": "openai-gpt-6-astra",
        "label": "openai/gpt-6-astra",
        "model": "openai/gpt-6-astra",
        "provider": "OpenAI",
        "harness": null,
        "effort": "max",
        "score": 63.125,
        "costUsd": 18.457813,
        "uncertainty": {
          "lower": 61.935,
          "upper": 64.315,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "openai/gpt-6-astra"
          },
          {
            "label": "Standard error",
            "value": "±1.19 points"
          },
          {
            "label": "Mean latency",
            "value": "1,668 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "53.5%"
          },
          {
            "label": "EMB",
            "value": "71.7%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "59.6%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "89.6%"
          },
          {
            "label": "Code Migration",
            "value": "68.5%"
          },
          {
            "label": "Legal Research",
            "value": "39.4%"
          },
          {
            "label": "HLAB",
            "value": "5.4%"
          },
          {
            "label": "Tax Agent",
            "value": "63.3%"
          }
        ]
      },
      {
        "id": "openai-gpt-6-luna",
        "label": "openai/gpt-6-luna",
        "model": "openai/gpt-6-luna",
        "provider": "OpenAI",
        "harness": null,
        "effort": "max",
        "score": 51.216,
        "costUsd": 0.431072,
        "uncertainty": {
          "lower": 50.144,
          "upper": 52.288000000000004,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "openai/gpt-6-luna"
          },
          {
            "label": "Standard error",
            "value": "±1.07 points"
          },
          {
            "label": "Mean latency",
            "value": "1,834 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "49.9%"
          },
          {
            "label": "EMB",
            "value": "68.5%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "13.6%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "81.6%"
          },
          {
            "label": "Code Migration",
            "value": "44.4%"
          },
          {
            "label": "Legal Research",
            "value": "30.3%"
          },
          {
            "label": "HLAB",
            "value": "2.9%"
          },
          {
            "label": "Tax Agent",
            "value": "58.9%"
          }
        ]
      },
      {
        "id": "openai-gpt-6-sol",
        "label": "openai/gpt-6-sol",
        "model": "openai/gpt-6-sol",
        "provider": "OpenAI",
        "harness": null,
        "effort": "max",
        "score": 57.536,
        "costUsd": 7.580816,
        "uncertainty": {
          "lower": 56.524,
          "upper": 58.548,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "openai/gpt-6-sol"
          },
          {
            "label": "Standard error",
            "value": "±1.01 points"
          },
          {
            "label": "Mean latency",
            "value": "1,751 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "49.0%"
          },
          {
            "label": "EMB",
            "value": "71.5%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "44.4%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "87.8%"
          },
          {
            "label": "Code Migration",
            "value": "56.9%"
          },
          {
            "label": "Legal Research",
            "value": "28.8%"
          },
          {
            "label": "HLAB",
            "value": "1.7%"
          },
          {
            "label": "Tax Agent",
            "value": "53.0%"
          }
        ]
      },
      {
        "id": "stepfun-step-5-preview",
        "label": "stepfun/step-5-preview",
        "model": "stepfun/step-5-preview",
        "provider": "Stepfun",
        "harness": null,
        "effort": null,
        "score": 49.354,
        "costUsd": 2.625676,
        "uncertainty": {
          "lower": 48.132,
          "upper": 50.576,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "stepfun/step-5-preview"
          },
          {
            "label": "Standard error",
            "value": "±1.22 points"
          },
          {
            "label": "Mean latency",
            "value": "4,218 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "50.7%"
          },
          {
            "label": "EMB",
            "value": "57.1%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "31.8%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "70.4%"
          },
          {
            "label": "Code Migration",
            "value": "39.5%"
          },
          {
            "label": "Legal Research",
            "value": "40.4%"
          },
          {
            "label": "HLAB",
            "value": "9.2%"
          },
          {
            "label": "Tax Agent",
            "value": "59.5%"
          }
        ]
      },
      {
        "id": "tencent-hy4-preview",
        "label": "tencent/hy4-preview",
        "model": "tencent/hy4-preview",
        "provider": "Tencent",
        "harness": null,
        "effort": null,
        "score": 49.942,
        "costUsd": 1.438575,
        "uncertainty": {
          "lower": 48.793,
          "upper": 51.091,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "tencent/hy4-preview"
          },
          {
            "label": "Standard error",
            "value": "±1.15 points"
          },
          {
            "label": "Mean latency",
            "value": "3,657 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "55.1%"
          },
          {
            "label": "EMB",
            "value": "58.1%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "8.1%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "77.5%"
          },
          {
            "label": "Code Migration",
            "value": "46.8%"
          },
          {
            "label": "Legal Research",
            "value": "45.2%"
          },
          {
            "label": "HLAB",
            "value": "9.2%"
          },
          {
            "label": "Tax Agent",
            "value": "63.7%"
          }
        ]
      },
      {
        "id": "thinkingmachines-inkling",
        "label": "thinkingmachines/inkling",
        "model": "thinkingmachines/inkling",
        "provider": "Thinkingmachines",
        "harness": null,
        "effort": "0.99",
        "score": 28.668,
        "costUsd": 1.464372,
        "uncertainty": {
          "lower": 27.567999999999998,
          "upper": 29.768,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "thinkingmachines/inkling"
          },
          {
            "label": "Standard error",
            "value": "±1.10 points"
          },
          {
            "label": "Mean latency",
            "value": "1,206 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "46.6%"
          },
          {
            "label": "EMB",
            "value": "38.4%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "0.5%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "19.2%"
          },
          {
            "label": "Code Migration",
            "value": "11.8%"
          },
          {
            "label": "Legal Research",
            "value": "28.4%"
          },
          {
            "label": "HLAB",
            "value": "2.1%"
          },
          {
            "label": "Tax Agent",
            "value": "43.5%"
          }
        ]
      },
      {
        "id": "thinkingmachines-inkling-small",
        "label": "thinkingmachines/inkling-small",
        "model": "thinkingmachines/inkling-small",
        "provider": "Thinkingmachines",
        "harness": null,
        "effort": "0.99",
        "score": 25.459,
        "costUsd": 0.281793,
        "uncertainty": {
          "lower": 24.427,
          "upper": 26.491,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "thinkingmachines/inkling-small"
          },
          {
            "label": "Standard error",
            "value": "±1.03 points"
          },
          {
            "label": "Mean latency",
            "value": "1,262 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "41.3%"
          },
          {
            "label": "EMB",
            "value": "31.8%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "0.0%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "19.1%"
          },
          {
            "label": "Code Migration",
            "value": "13.3%"
          },
          {
            "label": "Legal Research",
            "value": "25.5%"
          },
          {
            "label": "HLAB",
            "value": "1.7%"
          },
          {
            "label": "Tax Agent",
            "value": "41.5%"
          }
        ]
      },
      {
        "id": "xiaomi-mimo-v2-5-pro",
        "label": "xiaomi/mimo-v2.5-pro",
        "model": "xiaomi/mimo-v2.5-pro",
        "provider": "Xiaomi",
        "harness": null,
        "effort": null,
        "score": 33.842,
        "costUsd": 0.161806,
        "uncertainty": {
          "lower": 32.671,
          "upper": 35.013,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "xiaomi/mimo-v2.5-pro"
          },
          {
            "label": "Standard error",
            "value": "±1.17 points"
          },
          {
            "label": "Mean latency",
            "value": "2,073 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "41.5%"
          },
          {
            "label": "EMB",
            "value": "55.2%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "0.5%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "34.1%"
          },
          {
            "label": "Code Migration",
            "value": "20.4%"
          },
          {
            "label": "Legal Research",
            "value": "15.9%"
          },
          {
            "label": "HLAB",
            "value": "2.1%"
          },
          {
            "label": "Tax Agent",
            "value": "34.8%"
          }
        ]
      },
      {
        "id": "xiaomi-mimo-v2-6-flash",
        "label": "xiaomi/mimo-v2.6-flash",
        "model": "xiaomi/mimo-v2.6-flash",
        "provider": "Xiaomi",
        "harness": null,
        "effort": null,
        "score": 53.231,
        "costUsd": 0.202364,
        "uncertainty": {
          "lower": 52.134,
          "upper": 54.328,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "xiaomi/mimo-v2.6-flash"
          },
          {
            "label": "Standard error",
            "value": "±1.10 points"
          },
          {
            "label": "Mean latency",
            "value": "3,482 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "56.3%"
          },
          {
            "label": "EMB",
            "value": "65.5%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "24.2%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "79.0%"
          },
          {
            "label": "Code Migration",
            "value": "40.4%"
          },
          {
            "label": "Legal Research",
            "value": "38.0%"
          },
          {
            "label": "HLAB",
            "value": "11.3%"
          },
          {
            "label": "Tax Agent",
            "value": "59.9%"
          }
        ]
      },
      {
        "id": "xiaomi-mimo-v2-6-pro",
        "label": "xiaomi/mimo-v2.6-pro",
        "model": "xiaomi/mimo-v2.6-pro",
        "provider": "Xiaomi",
        "harness": null,
        "effort": null,
        "score": 55.199,
        "costUsd": 0.410427,
        "uncertainty": {
          "lower": 54.016999999999996,
          "upper": 56.381,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "xiaomi/mimo-v2.6-pro"
          },
          {
            "label": "Standard error",
            "value": "±1.18 points"
          },
          {
            "label": "Mean latency",
            "value": "4,039 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "57.3%"
          },
          {
            "label": "EMB",
            "value": "62.9%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "31.3%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "85.2%"
          },
          {
            "label": "Code Migration",
            "value": "42.3%"
          },
          {
            "label": "Legal Research",
            "value": "47.1%"
          },
          {
            "label": "HLAB",
            "value": "10.8%"
          },
          {
            "label": "Tax Agent",
            "value": "64.9%"
          }
        ]
      },
      {
        "id": "zai-glm-5-3",
        "label": "zai/glm-5.3",
        "model": "zai/glm-5.3",
        "provider": "Fireworks AI",
        "harness": null,
        "effort": "max",
        "score": 53.514,
        "costUsd": 7.248682,
        "uncertainty": {
          "lower": 52.209,
          "upper": 54.819,
          "label": "±1 standard error"
        },
        "sourceUrl": "https://www.vals.ai/benchmarks/vals_index",
        "details": [
          {
            "label": "Publisher model id",
            "value": "zai/glm-5.3"
          },
          {
            "label": "Standard error",
            "value": "±1.30 points"
          },
          {
            "label": "Mean latency",
            "value": "4,393 seconds per test"
          },
          {
            "label": "Evaluation mode",
            "value": "Agentic · many steps allowed"
          },
          {
            "label": "Finance Agent v2",
            "value": "55.8%"
          },
          {
            "label": "EMB",
            "value": "56.3%"
          },
          {
            "label": "Terminal-Bench 4.0",
            "value": "38.9%"
          },
          {
            "label": "Vibe Code Bench",
            "value": "78.1%"
          },
          {
            "label": "Code Migration",
            "value": "43.2%"
          },
          {
            "label": "Legal Research",
            "value": "49.0%"
          },
          {
            "label": "HLAB",
            "value": "8.3%"
          },
          {
            "label": "Tax Agent",
            "value": "73.1%"
          }
        ]
      }
    ]
  }
}