Acropolis

OSWorld

OSWorld results from the OSWorld site and vendor posts that Epoch AI republishes. The Agent column carries the step budget, for example `(100 steps)`. Scores are already 0 to 100.

Upstream source
AgenticVision
Independent. Run by an operator independent of the model vendors.Epoch AI Benchmarking Hub by Epoch AIVersion 1Data as of 17 Nov 2026

No filters. Showing 38 of 38 rows.

  1. 1
    AnthropicDossier

    Score

    Sonnet 4.6, Score: 72.1% (scale 0.0 to 100.0)72.1%
  2. 2
    AnthropicDossier

    Score

    Opus 4.5, Score: 66.3% (scale 0.0 to 100.0)66.3%
  3. 3
    Moonshot AIDossier

    Score

    Kimi K2.5, Score: 63.3% (scale 0.0 to 100.0)63.3%
  4. 4
    AnthropicDossier

    Score

    Sonnet 4.5, Score: 62.9% (scale 0.0 to 100.0)62.9%
  5. 5
    AnthropicDossier

    Score

    Sonnet 4.5, Score: 58.1% (scale 0.0 to 100.0)58.1%
  6. 6
    ByteDanceDossier

    Score

    UI-TARS-2 (2509), Score: 53.1% (scale 0.0 to 100.0)53.1%
  7. 7
    AnthropicDossier

    Score

    Claude Sonnet 4, Score: 43.9% (scale 0.0 to 100.0)43.9%
  8. 8
    AnthropicDossier

    Score

    Sonnet 4.5, Score: 42.9% (scale 0.0 to 100.0)42.9%
  9. 9
    ByteDanceDossier

    Score

    UI-TARS (250705), Score: 41.8% (scale 0.0 to 100.0)41.8%
  10. 10
    AlibabaDossier

    Score

    Qwen3 VL Flash, Score: 41.6% (scale 0.0 to 100.0)41.6%
  11. 11
    AnthropicDossier

    Score

    Claude Sonnet 4, Score: 41.4% (scale 0.0 to 100.0)41.4%
  12. 12
    ByteDanceDossier

    Score

    Doubao 1.5 Thinking Vision Pro (250717), Score: 40% (scale 0.0 to 100.0)40%
  13. 13
    AnthropicDossier

    Score

    Claude 3.7 Sonnet, Score: 35.8% (scale 0.0 to 100.0)35.8%
  14. 14
    AnthropicDossier

    Score

    Claude 3.7 Sonnet, Score: 35.6% (scale 0.0 to 100.0)35.6%
  15. 15
    ByteDanceDossier

    Score

    Doubao 1.5 Thinking Vision Pro (250428), Score: 33.8% (scale 0.0 to 100.0)33.8%
  16. 16
    AlibabaDossier

    Score

    GUI-Owl 7B, Score: 32.1% (scale 0.0 to 100.0)32.1%
  17. 17
    ByteDanceDossier

    Score

    Doubao 1.5 Thinking Vision Pro (250717), Score: 31.9% (scale 0.0 to 100.0)31.9%
  18. 18
    OpenAIDossier

    Score

    Computer Use Preview, Score: 31.3% (scale 0.0 to 100.0)31.3%
  19. 19
    AnthropicDossier

    Score

    Claude Sonnet 4, Score: 31.2% (scale 0.0 to 100.0)31.2%
  20. 20
    OpenAIDossier

    Score

    Computer Use Preview, Score: 30.5% (scale 0.0 to 100.0)30.5%
  21. 21
    ByteDanceDossier

    Score

    Doubao 1.5 Thinking Vision Pro (250428), Score: 27.8% (scale 0.0 to 100.0)27.8%
  22. 22
    ByteDanceDossier

    Score

    UI-TARS 1.5 7B, Score: 27.4% (scale 0.0 to 100.0)27.4%
  23. 23
    ByteDanceDossier

    Score

    UI-TARS 1.5 7B, Score: 27.3% (scale 0.0 to 100.0)27.3%
  24. 24
    AnthropicDossier

    Score

    Claude 3.7 Sonnet, Score: 27.1% (scale 0.0 to 100.0)27.1%
  25. 24
    ByteDanceDossier

    Score

    UI-TARS 72B DPO, Score: 27.1% (scale 0.0 to 100.0)27.1%
  26. 26
    OpenAIDossier

    Score

    Computer Use Preview, Score: 26% (scale 0.0 to 100.0)26%
  27. 27
    ByteDanceDossier

    Score

    UI-TARS 72B DPO, Score: 25.8% (scale 0.0 to 100.0)25.8%
  28. 28
    ByteDanceDossier

    Score

    UI-TARS 1.5 7B, Score: 24.5% (scale 0.0 to 100.0)24.5%
  29. 29
    ByteDanceDossier

    Score

    UI-TARS 72B DPO, Score: 24% (scale 0.0 to 100.0)24%
  30. 30
    OpenAIDossier

    medium

    Score

    o3, Score: 23% (scale 0.0 to 100.0)23%
  31. 31
    OpenAIDossier

    medium

    Score

    o3, Score: 17.2% (scale 0.0 to 100.0)17.2%
  32. 32
    Moonshot AIDossier

    Score

    Kimi-VL A3B, Score: 10.3% (scale 0.0 to 100.0)10.3%
  33. 33
    Moonshot AIDossier

    Score

    Kimi-VL A3B, Score: 9.7% (scale 0.0 to 100.0)9.7%
  34. 34
    OpenAIDossier

    medium

    Score

    o3, Score: 9.1% (scale 0.0 to 100.0)9.1%
  35. 35
    AlibabaDossier

    Score

    Qwen2.5 VL 72B Instruct, Score: 5% (scale 0.0 to 100.0)5%
  36. 36
    AlibabaDossier

    Score

    Qwen2.5 VL 72B Instruct, Score: 4.4% (scale 0.0 to 100.0)4.4%
  37. 37
    AlibabaDossier

    Score

    Qwen2.5 VL 32B Instruct, Score: 3.9% (scale 0.0 to 100.0)3.9%
  38. 38
    AlibabaDossier

    Score

    Qwen2.5 VL 32B Instruct, Score: 3% (scale 0.0 to 100.0)3%
Unmatched upstream rows excluded from this board (19)

These upstream rows could not be mapped cleanly to the canonical model registry or reasoning effort enum.

Raw nameParsed modelReasonDetail
agi-0 (50 steps)agi-0no-aliasscore
DeepMiner-Mano-72B (100 steps)DeepMiner-Mano-72Bno-aliasscore
opencua-72b-preview (100 steps)opencua-72b-previewno-aliasscore
opencua-72b-preview (50 steps)opencua-72b-previewno-aliasscore
DeepMiner-Mano-7B (100 steps)DeepMiner-Mano-7Bno-aliasscore
opencua-72b-preview (15 steps)opencua-72b-previewno-aliasscore
opencua-32b (100 steps)opencua-32bno-aliasscore
opencua-32b (50 steps)opencua-32bno-aliasscore
TianXi-Action-7B (50 steps)TianXi-Action-7Bno-aliasscore
opencua-32b (15 steps)opencua-32bno-aliasscore
opencua-7b (50 steps)opencua-7bno-aliasscore
opencua-7b (100 steps)opencua-7bno-aliasscore
opencua-7b (15 steps)opencua-7bno-aliasscore
opencua-qwen2-7b (100 steps)opencua-qwen2-7bno-aliasscore
opencua-qwen2-7b (50 steps)opencua-qwen2-7bno-aliasscore
opencua-a3b (50 steps)opencua-a3bno-aliasscore
opencua-qwen2-7b (15 steps)opencua-qwen2-7bno-aliasscore
opencua-a3b (100 steps)opencua-a3bno-aliasscore
opencua-a3b (15 steps)opencua-a3bno-aliasscore