{
  "schemaVersion": 1,
  "resultVersion": "v1",
  "generatedBy": "scripts/build-backrooms-results-table.mjs",
  "generatedAt": "2026-09-21T21:43:49.599Z",
  "dataUpdatedAt": "2026-09-20",
  "canonicalUrl": "https://benchmarks.bio/benchmarks/txbench-pp/",
  "benchmark": {
    "id": "txbench-pp",
    "label": "TxBench-Preclinical-Pharmacology",
    "category": "therapeutics",
    "horizon": "short",
    "evalCount": 100,
    "scoreScope": {
      "label": "Full benchmark",
      "evaluationCount": 100
    },
    "weight": 1,
    "description": "Agentic benchmark tasks for preclinical pharmacology analysis and decision-making.",
    "repository": "https://github.com/latchbio/txbench-pp",
    "paper": "https://latch.bio/txbench-pp",
    "paperPdfUrl": "https://benchmarks.bio/papers/txbench-pp.pdf",
    "interactiveUrl": "https://benchmarks.bio/txbench"
  },
  "scoreDefinition": "Published benchmark score on a 0–100 percentage scale using Full benchmark (100 evaluations); see the repository and paper for benchmark-specific grading methodology.",
  "results": [
    {
      "rank": 1,
      "id": "claude-opus-5-claude-code",
      "model": "Claude Opus 5",
      "harness": "Claude Code",
      "provider": "Anthropic",
      "scorePct": 68.67
    },
    {
      "rank": 2,
      "id": "grok-4-6-grok-build",
      "model": "Grok 4.6",
      "harness": "Grok Build",
      "provider": "SpaceXAI",
      "scorePct": 64.67
    },
    {
      "rank": 3,
      "id": "claude-opus-5-pi",
      "model": "Claude Opus 5",
      "harness": "Pi",
      "provider": "Anthropic",
      "scorePct": 63.33
    },
    {
      "rank": 4,
      "id": "grok-4-6-pi",
      "model": "Grok 4.6",
      "harness": "Pi",
      "provider": "SpaceXAI",
      "scorePct": 62
    },
    {
      "rank": 5,
      "id": "gpt-5-6-sol-pi",
      "model": "GPT-5.6 Sol",
      "harness": "Pi",
      "provider": "OpenAI",
      "scorePct": 61.67
    },
    {
      "rank": 6,
      "id": "gpt-6-astra-codex",
      "model": "GPT-6 Astra",
      "harness": "Codex",
      "provider": "OpenAI",
      "scorePct": 61.67
    },
    {
      "rank": 7,
      "id": "gpt-6-astra-pi",
      "model": "GPT-6 Astra",
      "harness": "Pi",
      "provider": "OpenAI",
      "scorePct": 60.33
    },
    {
      "rank": 8,
      "id": "claude-opus-4-8-pi",
      "model": "Claude Opus 4.8",
      "harness": "Pi",
      "provider": "Anthropic",
      "scorePct": 59.33
    },
    {
      "rank": 9,
      "id": "gpt-5-6-sol-codex",
      "model": "GPT-5.6 Sol",
      "harness": "Codex",
      "provider": "OpenAI",
      "scorePct": 59
    },
    {
      "rank": 10,
      "id": "grok-4-5-pi",
      "model": "Grok 4.5",
      "harness": "Pi",
      "provider": "SpaceXAI",
      "scorePct": 57.67
    },
    {
      "rank": 11,
      "id": "gpt-5-5-pi",
      "model": "GPT-5.5",
      "harness": "Pi",
      "provider": "OpenAI",
      "scorePct": 55.33
    },
    {
      "rank": 12,
      "id": "claude-opus-4-8-claude-code",
      "model": "Claude Opus 4.8",
      "harness": "Claude Code",
      "provider": "Anthropic",
      "scorePct": 54.67
    },
    {
      "rank": 13,
      "id": "deepseek-v4-1-flash-pi",
      "model": "DeepSeek V4.1 Flash",
      "harness": "Pi",
      "provider": "DeepSeek",
      "scorePct": 54.33
    },
    {
      "rank": 14,
      "id": "gpt-5-6-terra-pi",
      "model": "GPT-5.6 Terra",
      "harness": "Pi",
      "provider": "OpenAI",
      "scorePct": 54
    },
    {
      "rank": 15,
      "id": "gpt-5-6-terra-codex",
      "model": "GPT-5.6 Terra",
      "harness": "Codex",
      "provider": "OpenAI",
      "scorePct": 51.67
    },
    {
      "rank": 16,
      "id": "gemini-3-5-flash-pi",
      "model": "Gemini 3.5 Flash",
      "harness": "Pi",
      "provider": "Google",
      "scorePct": 51.33
    },
    {
      "rank": 17,
      "id": "claude-opus-4-7-pi",
      "model": "Claude Opus 4.7",
      "harness": "Pi",
      "provider": "Anthropic",
      "scorePct": 49.33
    },
    {
      "rank": 18,
      "id": "claude-sonnet-5-claude-code",
      "model": "Claude Sonnet 5",
      "harness": "Claude Code",
      "provider": "Anthropic",
      "scorePct": 48.33
    },
    {
      "rank": 19,
      "id": "gpt-5-5-codex",
      "model": "GPT-5.5",
      "harness": "Codex",
      "provider": "OpenAI",
      "scorePct": 47.33
    },
    {
      "rank": 20,
      "id": "gpt-5-6-luna-pi",
      "model": "GPT-5.6 Luna",
      "harness": "Pi",
      "provider": "OpenAI",
      "scorePct": 46.67
    },
    {
      "rank": 21,
      "id": "kimi-k3-pi",
      "model": "Kimi K3",
      "harness": "Pi",
      "provider": "Moonshot",
      "scorePct": 44.67
    },
    {
      "rank": 22,
      "id": "claude-opus-4-7-claude-code",
      "model": "Claude Opus 4.7",
      "harness": "Claude Code",
      "provider": "Anthropic",
      "scorePct": 44
    },
    {
      "rank": 23,
      "id": "claude-sonnet-5-pi",
      "model": "Claude Sonnet 5",
      "harness": "Pi",
      "provider": "Anthropic",
      "scorePct": 42
    },
    {
      "rank": 24,
      "id": "gpt-5-6-luna-codex",
      "model": "GPT-5.6 Luna",
      "harness": "Codex",
      "provider": "OpenAI",
      "scorePct": 42
    }
  ]
}