Acropolis

FrontierCode

Merge-quality coding score. Run by Cognition (a vendor); Epoch republishes the results. Aggregation per row, for example Mean@5.

Upstream source
CodingAgentic
Independent. Run by an operator independent of the model vendors.Epoch AI Benchmarking Hub by Epoch AIVersion unstatedData as of 17 Nov 2026

No filters. Showing 51 of 51 rows.

  1. 1
    AnthropicDossier

    medium · Claude Code

    Main score

    Opus 5.5, Main score: 54.6% (scale 0.0 to 100.0)54.6%
  2. 2
    AnthropicDossier

    xhigh · Claude Code

    Main score

    Fable 5, Main score: 53.5% (scale 0.0 to 100.0)53.5%
  3. 3
    AnthropicDossier

    medium · Claude Code

    Main score

    Opus 5, Main score: 53.4% (scale 0.0 to 100.0)53.4%
  4. 4
    OpenAIDossier

    max · Codex CLI

    Main score

    GPT-6 Astra, Main score: 53.3% (scale 0.0 to 100.0)53.3%
  5. 5
    AnthropicDossier

    xhigh · Claude Code

    Main score

    Sonnet 5.5, Main score: 52.1% (scale 0.0 to 100.0)52.1%
  6. 6
    AnthropicDossier

    medium · Claude Code

    Main score

    Fable 5.1, Main score: 50.9% (scale 0.0 to 100.0)50.9%
  7. 7
    OpenAIDossier

    medium · Codex CLI

    Main score

    GPT-6.1 Sol, Main score: 50.2% (scale 0.0 to 100.0)50.2%
  8. 8
    CognitionDossier

    max · devin

    Main score

    SWE-2, Main score: 50% (scale 0.0 to 100.0)50%
  9. 9
    OpenAIDossier

    max · Codex CLI

    Main score

    GPT-6 Sol, Main score: 49.3% (scale 0.0 to 100.0)49.3%
  10. 10

    high · grok-build

    Main score

    Grok 4.6, Main score: 48% (scale 0.0 to 100.0)48%
  11. 11

    high · grok-build

    Main score

    Grok 4.7, Main score: 47.6% (scale 0.0 to 100.0)47.6%
  12. 12
    GoogleDossier

    medium · devin-cli

    Main score

    Gemini 4 Argon, Main score: 47.6% (scale 0.0 to 100.0)47.6%
  13. 13
    OpenAIDossier

    max · Codex CLI

    Main score

    GPT-5.6 Sol, Main score: 47.5% (scale 0.0 to 100.0)47.5%
  14. 14
    AnthropicDossier

    max · Claude Code

    Main score

    Opus 4.8, Main score: 46.5% (scale 0.0 to 100.0)46.5%
  15. 15
    AnthropicDossier

    max · Claude Code

    Main score

    Haiku 5.5, Main score: 46.4% (scale 0.0 to 100.0)46.4%
  16. 16
    Moonshot AIDossier

    none · Mini-SWE-Agent

    Main score

    Kimi K3, Main score: 44.2% (scale 0.0 to 100.0)44.2%
  17. 17
    DeepSeekDossier

    high · devin-cli

    Main score

    DeepSeek V4.1 Flash, Main score: 43.9% (scale 0.0 to 100.0)43.9%
  18. 18
    GoogleDossier

    medium · devin-cli

    Main score

    Gemini 3.7 Flash, Main score: 43.6% (scale 0.0 to 100.0)43.6%
  19. 19
    OpenAIDossier

    xhigh · Codex CLI

    Main score

    GPT-5.5, Main score: 43% (scale 0.0 to 100.0)43%
  20. 20
    AnthropicDossier

    xhigh · Claude Code

    Main score

    Sonnet 5, Main score: 42.7% (scale 0.0 to 100.0)42.7%
  21. 21

    high · grok-build

    Main score

    Grok 4.5, Main score: 42.4% (scale 0.0 to 100.0)42.4%
  22. 22
    OpenAIDossier

    max · Codex CLI

    Main score

    GPT-6 Luna, Main score: 42.4% (scale 0.0 to 100.0)42.4%
  23. 23
    CognitionDossier

    none · devin

    Main score

    SWE-1.7, Main score: 42.3% (scale 0.0 to 100.0)42.3%
  24. 24
    CognitionDossier

    none · devin-cli

    Main score

    SWE-1.7, Main score: 42% (scale 0.0 to 100.0)42%
  25. 25
    OpenAIDossier

    max · Codex CLI

    Main score

    GPT-5.6 Terra, Main score: 41.3% (scale 0.0 to 100.0)41.3%
  26. 26
    GoogleDossier

    medium · devin-cli

    Main score

    Gemini 3.8 Flash, Main score: 41.2% (scale 0.0 to 100.0)41.2%
  27. 27

    max · devin-cli

    Main score

    GLM-5.3, Main score: 40.1% (scale 0.0 to 100.0)40.1%
  28. 28
    OpenAIDossier

    max · Codex CLI

    Main score

    GPT-5.6 Luna, Main score: 39.8% (scale 0.0 to 100.0)39.8%
  29. 29
    MistralDossier

    high · chisel

    Main score

    Mistral Large 4, Main score: 39.4% (scale 0.0 to 100.0)39.4%
  30. 30
    AnthropicDossier

    max · Claude Code

    Main score

    Opus 4.7, Main score: 38.5% (scale 0.0 to 100.0)38.5%
  31. 31
    GoogleDossier

    medium · devin-cli

    Main score

    Gemini 3.6 Flash, Main score: 34.4% (scale 0.0 to 100.0)34.4%
  32. 32

    max · devin-cli

    Main score

    GLM-5.3 Flash, Main score: 31.8% (scale 0.0 to 100.0)31.8%
  33. 33
    Moonshot AIDossier

    none · Mini-SWE-Agent

    Main score

    Kimi K2.7 Code, Main score: 30.1% (scale 0.0 to 100.0)30.1%
  34. 34
    DeepSeekDossier

    high · devin-cli

    Main score

    DeepSeek V4 Pro 0813, Main score: 28.6% (scale 0.0 to 100.0)28.6%
  35. 35
    OpenAIDossier

    xhigh · Codex CLI

    Main score

    GPT-5.4 mini, Main score: 27% (scale 0.0 to 100.0)27%
  36. 36
    AnthropicDossier

    high · Claude Code

    Main score

    Opus 4.6, Main score: 26.6% (scale 0.0 to 100.0)26.6%
  37. 37
    CursorDossier

    none · cursor-cli

    Main score

    Composer 2.5, Main score: 25.6% (scale 0.0 to 100.0)25.6%
  38. 38
    CursorDossier

    none · cursor-cli

    Main score

    Composer 2.5, Main score: 25.6% (scale 0.0 to 100.0)25.6%
  39. 39

    none · Mini-SWE-Agent

    Main score

    GLM-5.2, Main score: 24.5% (scale 0.0 to 100.0)24.5%
  40. 40
    AnthropicDossier

    max · Claude Code

    Main score

    Sonnet 4.6, Main score: 24.3% (scale 0.0 to 100.0)24.3%
  41. 41
    DeepSeekDossier

    high · devin-cli

    Main score

    DeepSeek V4-Flash 0731, Main score: 18.8% (scale 0.0 to 100.0)18.8%
  42. 42
    DeepSeekDossier

    none · Mini-SWE-Agent

    Main score

    DeepSeek V4-Pro, Main score: 17.6% (scale 0.0 to 100.0)17.6%
  43. 43
    MiniMaxDossier

    none · Mini-SWE-Agent

    Main score

    MiniMax M3, Main score: 14.7% (scale 0.0 to 100.0)14.7%
  44. 44
    Thinking MachinesDossier

    Mini-SWE-Agent

    Main score

    Inkling, Main score: 14% (scale 0.0 to 100.0)14%
  45. 44
    Thinking MachinesDossier

    Mini-SWE-Agent

    Main score

    Inkling, Main score: 14% (scale 0.0 to 100.0)14%
  46. 46
    NVIDIADossier

    none · devin-cli

    Main score

    Nemotron 3 Ultra 550B A55B, Main score: 13.6% (scale 0.0 to 100.0)13.6%
  47. 47
    AlibabaDossier

    none · Mini-SWE-Agent

    Main score

    Qwen3.7 Plus, Main score: 10.2% (scale 0.0 to 100.0)10.2%
  48. 47
    AlibabaDossier

    none · Mini-SWE-Agent

    Main score

    Qwen3.7 Plus, Main score: 10.2% (scale 0.0 to 100.0)10.2%
  49. 49
    CognitionDossier

    none · devin

    Main score

    SWE-1.6, Main score: 9.4% (scale 0.0 to 100.0)9.4%
  50. 50
    CognitionDossier

    none · devin-cli

    Main score

    SWE-1.6, Main score: 9.4% (scale 0.0 to 100.0)9.4%
  51. 51
    MistralDossier

    none · devin-cli

    Main score

    Mistral Medium (2604), Main score: 8% (scale 0.0 to 100.0)8%