Acropolis

Code Migration

Code Migration, a component of the Vals Index.

Upstream source
CodingAgentic
Independent. Run by an operator independent of the model vendors.Vals AI benchmarks by Vals AIVersion 1.1Data as of 17 Nov 2026

No filters. Showing 38 of 38 rows.

  1. 1
    AnthropicDossier

    max

    Accuracy

    Sonnet 5.5, Accuracy: 87.4% (scale 0.0 to 100.0), 95% confidence interval 82.7 to 92.187.4%(82.7–92.1)
    Duration per test
    14,850.4 s
    Cost per test
    $45.41
  2. 2
    AnthropicDossier

    high

    Tied #1

    Accuracy

    Fable 5.2, Accuracy: 85.8% (scale 0.0 to 100.0), 95% confidence interval 78.9 to 92.885.8%(78.9–92.8)
    Duration per test
    11,373.3 s
    Cost per test
    $68.68
  3. 3
    GoogleDossier

    xhigh

    Tied #1

    Accuracy

    Gemini 4 Hydrogen, Accuracy: 82.7% (scale 0.0 to 100.0), 95% confidence interval 76.7 to 88.682.7%(76.7–88.6)
    Duration per test
    7,150.7 s
    Cost per test
    $45.64
  4. 4
    AnthropicDossier

    max

    Tied #1

    Accuracy

    Opus 5.5, Accuracy: 80.5% (scale 0.0 to 100.0), 95% confidence interval 72.5 to 88.580.5%(72.5–88.5)
    Duration per test
    12,635.3 s
    Cost per test
    $72.30
  5. 5
    GoogleDossier

    high

    Tied #1

    Accuracy

    Gemini 4 Carbon, Accuracy: 80.1% (scale 0.0 to 100.0), 95% confidence interval 74.0 to 86.180.1%(74.0–86.1)
    Duration per test
    7,692.3 s
    Cost per test
    $33.71
  6. 6
    OpenAIDossier

    max

    Accuracy

    GPT-6 Astra, Accuracy: 78% (scale 0.0 to 100.0), 95% confidence interval 71.9 to 84.278%(71.9–84.2)
    Duration per test
    5,286.5 s
    Cost per test
    $66.64
  7. 7
    OpenAIDossier

    max

    Accuracy

    GPT-6.1 Sol, Accuracy: 75.1% (scale 0.0 to 100.0), 95% confidence interval 68.1 to 82.275.1%(68.1–82.2)
    Duration per test
    6,157.9 s
    Cost per test
    $9.94
  8. 8
    GoogleDossier

    high

    Accuracy

    Gemini 4 Argon, Accuracy: 74.4% (scale 0.0 to 100.0), 95% confidence interval 65.6 to 83.374.4%(65.6–83.3)
    Duration per test
    8,400.2 s
    Cost per test
    $44.64
  9. 9
    OpenAIDossier

    max

    Accuracy

    GPT-6.1 Terra, Accuracy: 72.2% (scale 0.0 to 100.0), 95% confidence interval 65.9 to 78.472.2%(65.9–78.4)
    Duration per test
    4,463.8 s
    Cost per test
    $6.02
  10. 10

    max

    Accuracy

    Muse Wonder 1.0, Accuracy: 69.8% (scale 0.0 to 100.0), 95% confidence interval 61.9 to 77.769.8%(61.9–77.7)
    Duration per test
    9,687.2 s
    Cost per test
    $47.55
  11. 11
    DeepSeekDossier

    max

    Accuracy

    DeepSeek V4.1 Flash 1017, Accuracy: 68.5% (scale 0.0 to 100.0), 95% confidence interval 61.7 to 75.368.5%(61.7–75.3)
    Duration per test
    8,133.1 s
    Cost per test
    $1.57
  12. 12

    max

    Accuracy

    Muse Wonder Full 1.0, Accuracy: 61.7% (scale 0.0 to 100.0), 95% confidence interval 53.1 to 70.261.7%(53.1–70.2)
    Duration per test
    15,566.7 s
    Cost per test
    $23.24
  13. 13
    Moonshot AIDossier

    max

    Accuracy

    Kimi K3.1 Pro, Accuracy: 59.7% (scale 0.0 to 100.0), 95% confidence interval 50.7 to 68.759.7%(50.7–68.7)
    Duration per test
    7,030.3 s
    Cost per test
    $15.71
  14. 14
    TencentDossier

    Accuracy

    Hy4, Accuracy: 59.1% (scale 0.0 to 100.0), 95% confidence interval 52.6 to 65.659.1%(52.6–65.6)
    Duration per test
    6,122.1 s
    Cost per test
    $3.24
  15. 15
    AlibabaDossier

    xhigh

    Accuracy

    Qwen4 Max, Accuracy: 57.3% (scale 0.0 to 100.0), 95% confidence interval 49.7 to 64.857.3%(49.7–64.8)
    Duration per test
    10,572.5 s
    Cost per test
    $10.11
  16. 16
    AlibabaDossier

    Accuracy

    Qwen3.8 Max Internal, Accuracy: 54.5% (scale 0.0 to 100.0), 95% confidence interval 47.4 to 61.754.5%(47.4–61.7)
    Duration per test
    14,749.8 s
    Cost per test
    $25.86
  17. 17

    Accuracy

    GLM-5.3, Accuracy: 54.1% (scale 0.0 to 100.0), 95% confidence interval 47.8 to 60.454.1%(47.8–60.4)
    Duration per test
    10,660.6 s
    Cost per test
    $33.52
  18. 18
    DeepSeekDossier

    max

    Accuracy

    DeepSeek V4 Pro 0813, Accuracy: 52% (scale 0.0 to 100.0), 95% confidence interval 45.4 to 58.652%(45.4–58.6)
    Duration per test
    10,883.5 s
    Cost per test
    $11.81
  19. 19

    xhigh

    Accuracy

    Grok 4.7, Accuracy: 51.8% (scale 0.0 to 100.0), 95% confidence interval 45.5 to 58.151.8%(45.5–58.1)
    Duration per test
    3,787.6 s
    Cost per test
    $32.98
  20. 20
    MistralDossier

    high

    Accuracy

    Mistral Large 4 OSS, Accuracy: 51.3% (scale 0.0 to 100.0), 95% confidence interval 42.6 to 60.151.3%(42.6–60.1)
    Duration per test
    10,868.3 s
    Cost per test
    $11.77
  21. 21

    Accuracy

    GLM-5.3 Flash, Accuracy: 50.6% (scale 0.0 to 100.0), 95% confidence interval 42.8 to 58.550.6%(42.8–58.5)
    Duration per test
    12,036.1 s
    Cost per test
    $8.27
  22. 22
    AnthropicDossier

    max

    Accuracy

    Haiku 5.5, Accuracy: 50.2% (scale 0.0 to 100.0), 95% confidence interval 42.0 to 58.450.2%(42.0–58.4)
    Duration per test
    5,148.2 s
    Cost per test
    $6.13
  23. 23
    MiniMaxDossier

    max

    Accuracy

    MiniMax M3.1 Flash, Accuracy: 48.2% (scale 0.0 to 100.0), 95% confidence interval 41.6 to 54.848.2%(41.6–54.8)
    Duration per test
    7,363.9 s
    Cost per test
    $1.73
  24. 24

    max

    Accuracy

    Muse Ember, Accuracy: 47.1% (scale 0.0 to 100.0), 95% confidence interval 39.0 to 55.247.1%(39.0–55.2)
    Duration per test
    4,439 s
    Cost per test
    $0.60
  25. 25
    XiaomiDossier

    Accuracy

    MiMo-V2.6-Pro, Accuracy: 47.1% (scale 0.0 to 100.0), 95% confidence interval 37.8 to 56.347.1%(37.8–56.3)
    Duration per test
    12,258.7 s
    Cost per test
    $0.89
  26. 26
    StepFunDossier

    Accuracy

    Step 5 Preview, Accuracy: 46.9% (scale 0.0 to 100.0), 95% confidence interval 39.3 to 54.646.9%(39.3–54.6)
    Duration per test
    11,290.2 s
    Cost per test
    $8.48
  27. 27
    OpenAIDossier

    max

    Accuracy

    GPT-6 Luna, Accuracy: 45.5% (scale 0.0 to 100.0), 95% confidence interval 37.2 to 53.745.5%(37.2–53.7)
    Duration per test
    3,543.6 s
    Cost per test
    $0.71
  28. 28
    TencentDossier

    Accuracy

    Hy4 Preview, Accuracy: 44.7% (scale 0.0 to 100.0), 95% confidence interval 36.1 to 53.344.7%(36.1–53.3)
    Duration per test
    6,574.3 s
    Cost per test
    $3.42
  29. 29
    Thinking MachinesDossier

    Accuracy

    Inkling 2 Small, Accuracy: 39.5% (scale 0.0 to 100.0), 95% confidence interval 30.3 to 48.639.5%(30.3–48.6)
    Duration per test
    13,105.7 s
    Cost per test
    $15.58
  30. 30
    XiaomiDossier

    Accuracy

    MiMo-V2.6-Flash, Accuracy: 38.8% (scale 0.0 to 100.0), 95% confidence interval 30.3 to 47.238.8%(30.3–47.2)
    Duration per test
    14,739.1 s
    Cost per test
    $0.61
  31. 31
    GoogleDossier

    high

    Accuracy

    Gemini 3.8 Flash, Accuracy: 38.2% (scale 0.0 to 100.0), 95% confidence interval 29.8 to 46.638.2%(29.8–46.6)
    Duration per test
    7,443.7 s
    Cost per test
    $8.31
  32. 32
    InceptionDossier

    high

    Accuracy

    Mercury 3, Accuracy: 36.9% (scale 0.0 to 100.0), 95% confidence interval 29.7 to 44.036.9%(29.7–44.0)
    Duration per test
    5,686.6 s
    Cost per test
    $2.21
  33. 33

    xhigh

    Accuracy

    Muse Spark Open, Accuracy: 25.1% (scale 0.0 to 100.0), 95% confidence interval 18.2 to 31.925.1%(18.2–31.9)
    Duration per test
    6,556.2 s
    Cost per test
    $4.64
  34. 34
    Moonshot AIDossier

    max

    Accuracy

    Kimi K3, Accuracy: 22.7% (scale 0.0 to 100.0), 95% confidence interval 15.7 to 29.622.7%(15.7–29.6)
    Duration per test
    16,267.5 s
    Cost per test
    $19.51
  35. 35
    ReflectionDossier

    Accuracy

    Beam 1.0, Accuracy: 22.6% (scale 0.0 to 100.0), 95% confidence interval 15.8 to 29.422.6%(15.8–29.4)
    Duration per test
    713.5 s
    Cost per test
    Not tested
  36. 36
    NVIDIADossier

    high

    Accuracy

    Nemotron 3.5 Super VL, Accuracy: 17.5% (scale 0.0 to 100.0), 95% confidence interval 11.3 to 23.717.5%(11.3–23.7)
    Duration per test
    1,729.6 s
    Cost per test
    $1.73
  37. 37
    Thinking MachinesDossier

    Accuracy

    Inkling Small, Accuracy: 10.4% (scale 0.0 to 100.0), 95% confidence interval 5.7 to 15.010.4%(5.7–15.0)
    Duration per test
    1,133.8 s
    Cost per test
    $1.59
  38. 38
    Thinking MachinesDossier

    Accuracy

    Inkling, Accuracy: 8.6% (scale 0.0 to 100.0), 95% confidence interval 3.2 to 14.08.6%(3.2–14.0)
    Duration per test
    1,414 s
    Cost per test
    $6.02