{
  "api_version": "1",
  "updated": "2026-10-16",
  "latest_event": "2026-10-15",
  "rights": "© Frontier Log, all rights reserved.",
  "benchmark": {
    "slug": "terminal-bench-science-0-1",
    "name": "Terminal-Bench Science 0.1",
    "family": "terminal-bench-science",
    "unit": "%",
    "higher_is_better": true,
    "status": "active",
    "benchmark_url": "https://www.tbench.ai",
    "settings": [
      {
        "setting": "default",
        "name": "",
        "url": "https://frontierlog.ai/benchmark/terminal-bench-science-0-1",
        "results": 3
      }
    ]
  },
  "total": 3,
  "results": [
    {
      "id": "gpt-6-astra:terminal-bench-science-0-1:default:1",
      "model": "gpt-6-astra",
      "model_name": "GPT-6 Astra",
      "model_date": "2026-09-03",
      "event": "https://frontierlog.ai/event/2026-09-03-gpt-6-astra",
      "benchmark": "terminal-bench-science-0-1",
      "setting": "default",
      "score": 64.6,
      "unit": "%",
      "printed": "64.6%",
      "reported_by": "openai",
      "label": "GPT‑6 Astra",
      "source": "https://openai.com/index/gpt-6-astra",
      "source_title": "GPT-6 Astra: A new generation of intelligence",
      "source_date": "2026-09-03",
      "where": "Academic table, row 'Terminal-Bench Science 0.1', column 'GPT‑6 Astra'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "",
      "human_verified": true,
      "human_verified_on": "2026-10-16"
    },
    {
      "id": "claude-opus-5-5:terminal-bench-science-0-1:default:1",
      "model": "claude-opus-5-5",
      "model_name": "Claude Opus 5.5",
      "model_date": "2026-09-22",
      "event": "https://frontierlog.ai/event/2026-09-22-claude-opus-5-5",
      "benchmark": "terminal-bench-science-0-1",
      "setting": "default",
      "score": 58.7,
      "unit": "%",
      "printed": "58.7%",
      "reported_by": "anthropic",
      "label": "Opus 5.5",
      "source": "https://www.anthropic.com/claude-opus-5-5",
      "source_title": "Introducing Claude Opus 5.5",
      "source_date": "2026-09-22",
      "where": "benchmark table, row 'Terminal-Bench-Science 0.1', column 'Opus 5.5'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "max effort; standard error ±3.5–5 pts",
      "human_verified": true,
      "human_verified_on": "2026-10-16"
    },
    {
      "id": "gemini-4-argon:terminal-bench-science-0-1:default:1",
      "model": "gemini-4-argon",
      "model_name": "Gemini 4 Argon",
      "model_date": "2026-09-30",
      "event": "https://frontierlog.ai/event/2026-09-30-gemini-4-argon",
      "benchmark": "terminal-bench-science-0-1",
      "setting": "default",
      "score": 57.6,
      "unit": "%",
      "printed": "57.6%",
      "reported_by": "google",
      "label": "Gemini 4 Argon",
      "source": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-4-argon/",
      "source_title": "Gemini 4 Argon: our next era of frontier intelligence",
      "source_date": "2026-09-30",
      "where": "benchmark table (image), row 'Terminal-Bench Science 0.1', column 'Gemini 4 Argon'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "",
      "human_verified": true,
      "human_verified_on": "2026-10-16",
      "source_image": "https://frontierlog.ai/sources/481de187c66d4b4d.gif",
      "source_image_original": "https://storage.googleapis.com/gweb-uniblog-publish-prod/original_images/gemini-4-argon_table_blog.gif"
    }
  ]
}