Vibeleaderboard
← Benchmark guide

Model results

GPQA Diamond

Choose a shared evaluation protocol to compare model scores. Versions, sources and configurations stay separate.

Atlas snapshot captured 10/19/2026, 10:20:28 AM UTC. This is an import timestamp, not the evaluation date. Coverage may be incomplete.

Reviewed recommendation and limits

Use as a compact reference for expert-written science questions, alongside evidence that matches the scientific work you intend to do.

Original GPQA Diamond subset. Preserve prompt, answer extraction, tools, sampling and aggregation. Keep main, extended and third-party corrected subsets separate.. Imported groups below may use different protocols; this review does not certify each run. The static 198-question subset has limited statistical power and exposure risk. Its expert filtering does not make it representative of scientific practice. A Google-proof collection process does not prove present-day contamination resistance.

Publisher results GPQA paper
← All evaluation groups

GPQA Diamond

Source
Epoch AI Benchmarking Hub
Version
Not reported
Measurement date
Not reported
Metric
% · Higher is better
Evaluation scope
Epoch AI run · task v1 · effort low · result date not supplied

Exact score ties share a position. Positions do not establish statistical significance. Missing metadata is shown as unreported; matching imported metadata is not independent verification of the protocol.

  1. #1GPT-5.5

    openai/gpt-5.5-20260423

    90.7%

    View modelEvidence independent

  2. #2GPT-5.6 Sol

    openai/gpt-5.6-sol-20260709

    89.9%

    View modelEvidence independent

  3. #3Gemini 3.5 Flash

    google/gemini-3.5-flash-20260519

    88.9%

    View modelEvidence independent

  4. #4Claude Opus 4.8

    anthropic/claude-opus-4-8

    88.4%

    View modelEvidence independent

  5. #5Claude Opus 5

    anthropic/claude-opus-5

    87.9%

    View modelEvidence independent

  6. #5GLM 5.2

    z-ai/glm-5.2-20260616

    87.9%

    View modelEvidence independent

  7. #7GPT-5.6 Terra

    openai/gpt-5.6-terra-20260709

    87.4%

    View modelEvidence independent

  8. #8Gemini 3.6 Flash

    google/gemini-3.6-flash-20260721

    86.4%

    View modelEvidence independent

  9. #9GPT-5.4

    openai/gpt-5.4-20260305

    84.8%

    View modelEvidence independent

  10. #9Kimi K3

    moonshotai/kimi-k3-20260715

    84.8%

    View modelEvidence independent

  11. #11GPT-5.2

    openai/gpt-5.2-20251211

    82.7%

    View modelEvidence independent

  12. #12GPT-5.6 Luna

    openai/gpt-5.6-luna-20260709

    82.3%

    View modelEvidence independent

  13. #13o3

    openai/o3-2025-04-16

    79.8%

    View modelEvidence independent

  14. #14Claude Fable 5

    anthropic/claude-fable-5

    78.8%

    View modelEvidence independent

  15. #15Gemini 3.5 Flash Lite

    google/gemini-3.5-flash-lite-20260721

    75.8%

    View modelEvidence independent

  16. #16o4 Mini

    openai/o4-mini-2025-04-16

    75.3%

    View modelEvidence independent

  17. #17Gemini 3.1 Flash Lite

    google/gemini-3.1-flash-lite-20260507

    74.2%

    View modelEvidence independent

  18. #17o1

    openai/o1-2024-12-17

    74.2%

    View modelEvidence independent

  19. #19GPT-5.4 Nano

    openai/gpt-5.4-nano-20260317

    72.2%

    View modelEvidence independent

  20. #20o3 Mini

    openai/o3-mini-2025-01-31

    68.2%

    View modelEvidence independent

  21. #21GPT-5 Nano

    openai/gpt-5-nano-2025-08-07

    57.6%

    View modelEvidence independent

  22. #22gpt-oss-20b

    openai/gpt-oss-20b

    53.2%

    View modelEvidence independent
GPQA Diamond model results | VibeLeaderboard