Vibeleaderboard
← Benchmark guide

Model results

FrontierMath

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 the relevant tier and private/public split for difficult mathematical reasoning.

Each tier, release and public/private split separately. Imported groups below may use different protocols; this review does not certify each run. Difficulty is specialized; tier, dataset release, tools and grader must match before comparing scores.

Publisher results Epoch benchmarking methods
← All evaluation groups

FrontierMath Tier 4 v2 · private

Source
Epoch AI Benchmarking Hub
Version
tier-4-v2-private
Measurement date
Not reported
Metric
% · Higher is better
Evaluation scope
Epoch AI run · task v2 · effort max · 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-6.1 Sol

    openai/gpt-6.1-sol-20260929

    100%

    View modelEvidence independent

  2. #2GPT-6 Astra

    openai/gpt-6-astra-20260903

    97.6%

    View modelEvidence independent

  3. #3Claude Opus 5.5

    anthropic/claude-opus-5-5

    95%

    View modelEvidence independent

  4. #4Claude Fable 5

    anthropic/claude-fable-5

    90.2%

    View modelEvidence independent

  5. #5GPT-6 Sol

    openai/gpt-6-sol-20260922

    90%

    View modelEvidence independent

  6. #6Claude Fable 5.1

    anthropic/claude-fable-5-1

    87.8%

    View modelEvidence independent

  7. #7GPT-5.6 Sol

    openai/gpt-5.6-sol-20260709

    82.9%

    View modelEvidence independent

  8. #8Claude Sonnet 5.5

    anthropic/claude-sonnet-5-5

    80.5%

    View modelEvidence independent

  9. #9Claude Opus 5

    anthropic/claude-opus-5

    73.2%

    View modelEvidence independent

  10. #10GPT-5.6 Terra

    openai/gpt-5.6-terra-20260709

    70.7%

    View modelEvidence independent

  11. #11GPT-5.6 Luna

    openai/gpt-5.6-luna-20260709

    61%

    View modelEvidence independent

  12. #12Claude Opus 4.8

    anthropic/claude-opus-4-8

    56.1%

    View modelEvidence independent

  13. #12GPT-6 Luna

    openai/gpt-6-luna-20260922

    56.1%

    View modelEvidence independent

  14. #14Claude Haiku 5.5

    anthropic/claude-haiku-5-5

    46.3%

    View modelEvidence independent

  15. #14Muse Spark 1.3

    meta/muse-spark-1.3-20260902

    46.3%

    View modelEvidence independent

  16. #16Kimi K3

    moonshotai/kimi-k3-20260715

    39%

    View modelEvidence independent

  17. #17Claude Opus 4.7

    anthropic/claude-opus-4-7

    31.7%

    View modelEvidence independent

  18. #18Claude Sonnet 5

    anthropic/claude-sonnet-5

    29.3%

    View modelEvidence independent

  19. #18GLM 5.2

    z-ai/glm-5.2-20260616

    29.3%

    View modelEvidence independent

  20. #18GLM 5.3

    z-ai/glm-5.3-20260816

    29.3%

    View modelEvidence independent

  21. #21Claude Opus 4.6

    anthropic/claude-opus-4-6

    26.8%

    View modelEvidence independent

  22. #21DeepSeek V4 Pro 0813

    deepseek/deepseek-v4-pro-20260813

    26.8%

    View modelEvidence independent

  23. #21DeepSeek V4.1 Flash

    deepseek/deepseek-v4.1-flash-20260910

    26.8%

    View modelEvidence independent

  24. #24DeepSeek V4 Flash 0731

    deepseek/deepseek-v4-flash-20260731

    24.4%

    View modelEvidence independent

  25. #25GLM 5.3 Flash

    z-ai/glm-5.3-flash-20260826

    17.1%

    View modelEvidence independent

  26. #26DeepSeek V4 Pro 0423

    deepseek/deepseek-v4-pro-20260423

    2.4%

    View modelEvidence independent
FrontierMath model results | VibeLeaderboard