Acropolis

Finance Agent

Finance Agent, a component of the Vals Index.

Upstream source
FinanceAgentic
Independent. Run by an operator independent of the model vendors.Vals AI benchmarks by Vals AIVersion 2.1Data as of 17 Nov 2026

No filters. Showing 38 of 38 rows.

  1. 1
    AnthropicDossier

    high

    Accuracy

    Fable 5.2, Accuracy: 69.5% (scale 0.0 to 100.0), 95% confidence interval 65.6 to 73.369.5%(65.6–73.3)
    Duration per test
    207.6 s
    Cost per test
    $4.83
  2. 2
    AnthropicDossier

    max

    Tied #1

    Accuracy

    Opus 5.5, Accuracy: 67.8% (scale 0.0 to 100.0), 95% confidence interval 63.8 to 71.967.8%(63.8–71.9)
    Duration per test
    1,980.1 s
    Cost per test
    $8.91
  3. 3
    GoogleDossier

    high

    Tied #1

    Accuracy

    Gemini 4 Argon, Accuracy: 67.7% (scale 0.0 to 100.0), 95% confidence interval 63.7 to 71.667.7%(63.7–71.6)
    Duration per test
    793 s
    Cost per test
    $4.32
  4. 4

    max

    Tied #1

    Accuracy

    Muse Wonder Full 1.0, Accuracy: 67.6% (scale 0.0 to 100.0), 95% confidence interval 63.7 to 71.567.6%(63.7–71.5)
    Duration per test
    850.3 s
    Cost per test
    $2.68
  5. 5
    Thinking MachinesDossier
    Tied #1

    Accuracy

    Inkling 2 Small, Accuracy: 67.5% (scale 0.0 to 100.0), 95% confidence interval 63.6 to 71.567.5%(63.6–71.5)
    Duration per test
    296.3 s
    Cost per test
    $0.18
  6. 6
    AnthropicDossier

    max

    Tied #1

    Accuracy

    Sonnet 5.5, Accuracy: 67.3% (scale 0.0 to 100.0), 95% confidence interval 63.3 to 71.267.3%(63.3–71.2)
    Duration per test
    2,066.2 s
    Cost per test
    $6.95
  7. 7
    GoogleDossier

    high

    Tied #1

    Accuracy

    Gemini 4 Carbon, Accuracy: 67.2% (scale 0.0 to 100.0), 95% confidence interval 63.2 to 71.167.2%(63.2–71.1)
    Duration per test
    1,779.6 s
    Cost per test
    $7.29
  8. 8
    AlibabaDossier

    xhigh

    Tied #1

    Accuracy

    Qwen4 Max, Accuracy: 66.7% (scale 0.0 to 100.0), 95% confidence interval 62.7 to 70.666.7%(62.7–70.6)
    Duration per test
    1,465.3 s
    Cost per test
    $1.74
  9. 9
    GoogleDossier

    high

    Tied #1

    Accuracy

    Gemini 3.8 Flash, Accuracy: 65.8% (scale 0.0 to 100.0), 95% confidence interval 61.8 to 69.865.8%(61.8–69.8)
    Duration per test
    443.9 s
    Cost per test
    $1.99
  10. 10

    max

    Tied #1

    Accuracy

    Muse Wonder 1.0, Accuracy: 65.5% (scale 0.0 to 100.0), 95% confidence interval 61.6 to 69.565.5%(61.6–69.5)
    Duration per test
    310.1 s
    Cost per test
    $2.62
  11. 11
    AnthropicDossier

    max

    Tied #1

    Accuracy

    Haiku 5.5, Accuracy: 65.4% (scale 0.0 to 100.0), 95% confidence interval 61.3 to 69.465.4%(61.3–69.4)
    Duration per test
    1,864.6 s
    Cost per test
    $1.64
  12. 12
    GoogleDossier

    xhigh

    Tied #1

    Accuracy

    Gemini 4 Hydrogen, Accuracy: 65.1% (scale 0.0 to 100.0), 95% confidence interval 61.1 to 69.165.1%(61.1–69.1)
    Duration per test
    1,078.1 s
    Cost per test
    $4.36
  13. 13

    max

    Tied #1

    Accuracy

    GLM-5.3, Accuracy: 65% (scale 0.0 to 100.0), 95% confidence interval 61.1 to 68.965%(61.1–68.9)
    Duration per test
    1,205.6 s
    Cost per test
    $1.03
  14. 14
    TencentDossier
    Tied #1

    Accuracy

    Hy4, Accuracy: 64.7% (scale 0.0 to 100.0), 95% confidence interval 60.7 to 68.764.7%(60.7–68.7)
    Duration per test
    2,238.9 s
    Cost per test
    $1.28
  15. 15
    XiaomiDossier
    Tied #1

    Accuracy

    MiMo-V2.6-Pro, Accuracy: 64.3% (scale 0.0 to 100.0), 95% confidence interval 60.3 to 68.264.3%(60.3–68.2)
    Duration per test
    682 s
    Cost per test
    $0.18
  16. 16

    max

    Accuracy

    Muse Ember, Accuracy: 63.8% (scale 0.0 to 100.0), 95% confidence interval 59.9 to 67.863.8%(59.9–67.8)
    Duration per test
    89.4 s
    Cost per test
    $0.056
  17. 17
    XiaomiDossier

    Accuracy

    MiMo-V2.6-Flash, Accuracy: 63.3% (scale 0.0 to 100.0), 95% confidence interval 59.2 to 67.463.3%(59.2–67.4)
    Duration per test
    936.3 s
    Cost per test
    $0.067
  18. 18
    DeepSeekDossier

    max

    Accuracy

    DeepSeek V4.1 Flash 1017, Accuracy: 62.9% (scale 0.0 to 100.0), 95% confidence interval 59.0 to 66.962.9%(59.0–66.9)
    Duration per test
    909.2 s
    Cost per test
    $0.47
  19. 19
    OpenAIDossier

    max

    Accuracy

    GPT-6.1 Sol, Accuracy: 61.8% (scale 0.0 to 100.0), 95% confidence interval 57.7 to 66.061.8%(57.7–66.0)
    Duration per test
    1,145.1 s
    Cost per test
    $1.63
  20. 20
    MistralDossier

    high

    Accuracy

    Mistral Large 4 OSS, Accuracy: 61.7% (scale 0.0 to 100.0), 95% confidence interval 57.7 to 65.761.7%(57.7–65.7)
    Duration per test
    387.2 s
    Cost per test
    $0.35
  21. 21
    TencentDossier

    Accuracy

    Hy4 Preview, Accuracy: 61.6% (scale 0.0 to 100.0), 95% confidence interval 57.6 to 65.761.6%(57.6–65.7)
    Duration per test
    1,502.7 s
    Cost per test
    $0.67
  22. 22
    Moonshot AIDossier

    max

    Accuracy

    Kimi K3.1 Pro, Accuracy: 61.5% (scale 0.0 to 100.0), 95% confidence interval 57.4 to 65.661.5%(57.4–65.6)
    Duration per test
    2,099 s
    Cost per test
    $2.29
  23. 23

    xhigh

    Accuracy

    Muse Spark Open, Accuracy: 60.7% (scale 0.0 to 100.0), 95% confidence interval 56.6 to 64.760.7%(56.6–64.7)
    Duration per test
    1,612.3 s
    Cost per test
    $2.76
  24. 24

    max

    Accuracy

    GLM-5.3 Flash, Accuracy: 60.6% (scale 0.0 to 100.0), 95% confidence interval 56.6 to 64.660.6%(56.6–64.6)
    Duration per test
    296.1 s
    Cost per test
    $0.072
  25. 25

    xhigh

    Accuracy

    Grok 4.7, Accuracy: 60.6% (scale 0.0 to 100.0), 95% confidence interval 56.4 to 64.760.6%(56.4–64.7)
    Duration per test
    652.3 s
    Cost per test
    $2.24
  26. 26
    OpenAIDossier

    max

    Accuracy

    GPT-6 Astra, Accuracy: 60.4% (scale 0.0 to 100.0), 95% confidence interval 56.3 to 64.660.4%(56.3–64.6)
    Duration per test
    1,228.8 s
    Cost per test
    $7.82
  27. 27
    Moonshot AIDossier

    max

    Accuracy

    Kimi K3, Accuracy: 60% (scale 0.0 to 100.0), 95% confidence interval 55.9 to 64.160%(55.9–64.1)
    Duration per test
    875 s
    Cost per test
    $1.88
  28. 28
    StepFunDossier

    Accuracy

    Step 5 Preview, Accuracy: 58.9% (scale 0.0 to 100.0), 95% confidence interval 54.9 to 62.958.9%(54.9–62.9)
    Duration per test
    542.5 s
    Cost per test
    $0.58
  29. 29
    AlibabaDossier

    Accuracy

    Qwen3.8 Max Internal, Accuracy: 58.8% (scale 0.0 to 100.0), 95% confidence interval 54.7 to 63.058.8%(54.7–63.0)
    Duration per test
    3,319.6 s
    Cost per test
    $2.99
  30. 30
    MiniMaxDossier

    max

    Accuracy

    MiniMax M3.1 Flash, Accuracy: 58.8% (scale 0.0 to 100.0), 95% confidence interval 54.8 to 62.958.8%(54.8–62.9)
    Duration per test
    904.2 s
    Cost per test
    $0.16
  31. 31
    DeepSeekDossier

    max

    Accuracy

    DeepSeek V4 Pro 0813, Accuracy: 57.1% (scale 0.0 to 100.0), 95% confidence interval 53.0 to 61.257.1%(53.0–61.2)
    Duration per test
    477 s
    Cost per test
    $0.82
  32. 32
    OpenAIDossier

    max

    Accuracy

    GPT-6 Luna, Accuracy: 56.8% (scale 0.0 to 100.0), 95% confidence interval 52.7 to 61.056.8%(52.7–61.0)
    Duration per test
    926.9 s
    Cost per test
    $0.12
  33. 33
    OpenAIDossier

    max

    Accuracy

    GPT-6.1 Terra, Accuracy: 56% (scale 0.0 to 100.0), 95% confidence interval 51.9 to 60.156%(51.9–60.1)
    Duration per test
    1,787.1 s
    Cost per test
    $1.92
  34. 34
    InceptionDossier

    high

    Accuracy

    Mercury 3, Accuracy: 54.7% (scale 0.0 to 100.0), 95% confidence interval 50.8 to 58.654.7%(50.8–58.6)
    Duration per test
    232.8 s
    Cost per test
    $0.065
  35. 35
    ReflectionDossier

    Accuracy

    Beam 1.0, Accuracy: 51.8% (scale 0.0 to 100.0), 95% confidence interval 47.7 to 56.051.8%(47.7–56.0)
    Duration per test
    85.4 s
    Cost per test
    Not tested
  36. 36
    Thinking MachinesDossier

    Accuracy

    Inkling, Accuracy: 49.9% (scale 0.0 to 100.0), 95% confidence interval 45.7 to 54.049.9%(45.7–54.0)
    Duration per test
    573.7 s
    Cost per test
    $0.89
  37. 37
    NVIDIADossier

    high

    Accuracy

    Nemotron 3.5 Super VL, Accuracy: 47.7% (scale 0.0 to 100.0), 95% confidence interval 43.6 to 51.847.7%(43.6–51.8)
    Duration per test
    79.5 s
    Cost per test
    $0.11
  38. 38
    Thinking MachinesDossier

    Accuracy

    Inkling Small, Accuracy: 44.8% (scale 0.0 to 100.0), 95% confidence interval 40.7 to 48.944.8%(40.7–48.9)
    Duration per test
    314.2 s
    Cost per test
    $0.21