Benchmark Almanac

BAAnchmark

Claude Sonnet 5.5 1.00 leads since September 2026
Not in model cards

Models build a 3D sheep herding game from the same Slovak prompt. Marek Dubovsky plays each runnable game and ranks its controls, playability, and bugs.

Dubovsky ranks 68 runs and publishes playable games. The results identify at least 41 distinct model IDs and document follow-up prompts for individual runs.

Made by
Dubovsky Labs
Kind
Web and app front ends
Level
University
Scored
Graded by people
Answers
Code
Tasks
All public
Measures
Capability
Language
Slovak
In and out
Text → Code
Interaction
Multi-turn
Items from
Written by experts
Refresh
One fixed set
Run by
Anyone
Strongest model tested
Claude Fable 5.2 #2 on the capability index
Last model added
November 2026
Scores refreshed
November 20th, 2026
62
Entries on its own leaderboard, 57 models we track

Claude Sonnet 5.5 leads with 1.00 Claude Sonnet 5.5 leads with 1.00

Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Every run is in the next chart. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. This view shows only models we track, each at its best.

  1. 1 Claude Sonnet 5.5 1.00
  2. 2 Claude Opus 5.5 multi-prompt 2.00
  3. 3 Claude Fable 5.2 3.00
  4. 4 Claude Opus 5 4.00
  5. 5 Gemini 4 Carbon 5.00
  6. 6 Claude Fable 5.1 one-shot 6.00
  7. 7 Claude Fable 5 7.00
  8. 8 DeepSeek V4.1 Flash 1017 8.00
  9. 9 Kimi K3 9.00
  10. 10 GLM-5.3-Flash multi-prompt 10.0
  11. 11 GPT-6 Astra one-shot 11.0
  12. 12 Union Alpha multi-prompt 12.0
  13. 13 Qwen3.8-Flash-Next Q2, multi-prompt 13.0
  14. 14 DeepSeek V4.1 Flash one-shot 14.0
  15. 15 Gemini 4 Argon 15.0
  16. 16 Qwen 3.8 27B ThinkingCap, multi-prompt 16.0
  17. 17 GPT-6 Sol 17.0
  18. 18 GPT-6.1 Sol 18.0
  19. 19 GPT-5.6 Sol 19.0
  20. 20 Claude Opus 4.8 20.0
  21. 21 Muse Spark 1.3 one-shot 21.0
  22. 22 Grok 4.7 22.0
  23. 23 Grok 4.6 one-shot 23.0
  24. 24 Grok 4.5 24.0
  25. 25 Gemini 3.8 Flash one-shot 25.0
  26. 26 Gemini 3.5 Flash 26.0
  27. 27 GPT-6 Luna multi-prompt 27.0
  28. 28 Gemini 3.6 Flash 29.0
  29. 29 Qwen 3.6 27B ThinkingCap ThinkingCap 30.0
  30. 30 Muse Spark 1.2 31.0
  31. 31 GPT-5.6 Terra one-shot 32.0
  32. 32 Qwen3.5 397B-A17B 33.0
  33. 33 MiniMax-M2.7 34.0
  34. 34 Qwen 3.8 Max 35.0
  35. 35 DeepSeek-V4-Pro 0831, multi-prompt 36.0
  36. 36 DeepSeek V4 Flash 0731 DS4, one-shot 37.0
  37. 37 Gemini 3.7 Flash multi-prompt 40.0
  38. 38 DeepSeek-V4-Flash-Vision-Exp multi-prompt 41.0
  39. 39 GPT-5.5 42.0
  40. 40 MiniMax-M3 43.0
  41. 41 Claude Sonnet 4.6 44.0
  42. 42 GLM-5.2 45.0
  43. 43 MiMo-V2.6 Pro 46.0
  44. 44 MiMo-V2.6 Flash 47.0
  45. 45 Mistral Large 4 48.0
  46. 46 DeepSeek-V4-Flash 49.0
  47. 47 Qwen3.7-Max 50.0
  48. 48 Big Pickle 52.0
  49. 49 Kimi K2.7 Code 53.0
  50. 50 Qwen3.6 27B MLX, multi-prompt 54.0
  51. 51 Qwen3.5-122B-A10B 56.0
  52. 52 glm-4.7-flash 57.0
  53. 53 Qwen 3.6 35B-A3B 58.0
  54. 54 Muse Glimmer one-shot 59.0
  55. 55 Gemma 4 31B IT 60.0
  56. 56 Laguna S 2.1 multi-prompt 61.0
  57. 57 GPT-5.6 Luna one-shot 62.0
  58. 58 MiMo v5 63.0
  59. 59 Mistral Small 4 64.0
  60. 60 Ornith 1.0 35B 65.0
  61. 61 Qwen3.5-35B-A3B 66.0
  62. 62 Nemotron 3 Ultra 68.0

Source: Official BAAnchmark leaderboard

  1. 1 Claude Sonnet 5.5 1.00
  2. 2 Claude Opus 5.5 multi-prompt 2.00
  3. 3 Claude Fable 5.2 3.00
  4. 4 Claude Opus 5 4.00
  5. 5 Gemini 4 Carbon 5.00
  6. 6 Claude Fable 5.1 one-shot 6.00
  7. 7 Claude Fable 5 7.00
  8. 8 DeepSeek V4.1 Flash 1017 8.00
  9. 9 Kimi K3 9.00
  10. 10 GLM-5.3-Flash multi-prompt 10.0
  11. 11 GPT-6 Astra one-shot 11.0
  12. 12 Qwen3.8-Flash-Next Q2, multi-prompt 13.0
  13. 13 DeepSeek V4.1 Flash one-shot 14.0
  14. 14 Gemini 4 Argon 15.0
  15. 15 Qwen 3.8 27B ThinkingCap, multi-prompt 16.0
  16. 16 GPT-6 Sol 17.0
  17. 17 GPT-6.1 Sol 18.0
  18. 18 GPT-5.6 Sol 19.0
  19. 19 Claude Opus 4.8 20.0
  20. 20 Muse Spark 1.3 one-shot 21.0
  21. 21 Grok 4.7 22.0
  22. 22 Grok 4.6 one-shot 23.0
  23. 23 Grok 4.5 24.0
  24. 24 Gemini 3.8 Flash one-shot 25.0
  25. 25 Gemini 3.5 Flash 26.0
  26. 26 GPT-6 Luna multi-prompt 27.0
  27. 27 Gemini 3.6 Flash 29.0
  28. 28 Muse Spark 1.2 31.0
  29. 29 GPT-5.6 Terra one-shot 32.0
  30. 30 Qwen3.5 397B-A17B 33.0
  31. 31 MiniMax-M2.7 34.0
  32. 32 Qwen 3.8 Max 35.0
  33. 33 DeepSeek-V4-Pro 0831, multi-prompt 36.0
  34. 34 DeepSeek V4 Flash 0731 DS4, one-shot 37.0
  35. 35 Gemini 3.7 Flash multi-prompt 40.0
  36. 36 DeepSeek-V4-Flash-Vision-Exp multi-prompt 41.0
  37. 37 GPT-5.5 42.0
  38. 38 MiniMax-M3 43.0
  39. 39 Claude Sonnet 4.6 44.0
  40. 40 GLM-5.2 45.0
  41. 41 MiMo-V2.6 Pro 46.0
  42. 42 MiMo-V2.6 Flash 47.0
  43. 43 Mistral Large 4 48.0
  44. 44 DeepSeek-V4-Flash 49.0
  45. 45 Qwen3.7-Max 50.0
  46. 46 Kimi K2.7 Code 53.0
  47. 47 Qwen3.6 27B MLX, multi-prompt 54.0
  48. 48 Qwen3.5-122B-A10B 56.0
  49. 49 glm-4.7-flash 57.0
  50. 50 Qwen 3.6 35B-A3B 58.0
  51. 51 Muse Glimmer one-shot 59.0
  52. 52 Gemma 4 31B IT 60.0
  53. 53 Laguna S 2.1 multi-prompt 61.0
  54. 54 GPT-5.6 Luna one-shot 62.0
  55. 55 Mistral Small 4 64.0
  56. 56 Qwen3.5-35B-A3B 66.0
  57. 57 Nemotron 3 Ultra 68.0

Source: Official BAAnchmark leaderboard

#EntryScoreSettingReleasedRun onLinkSource
1 Claude Sonnet 5.5 1.00 Sep 2026 Oct 2026 dubovskylabs.com Official leaderboard
2 Claude Opus 5.5 2.00 multi-prompt Sep 2026 Sep 2026 dubovskylabs.com Official leaderboard
3 Claude Fable 5.2 3.00 Nov 2026 Nov 2026 dubovskylabs.com Official leaderboard
4 Claude Opus 5 4.00 Jul 2026 dubovskylabs.com Official leaderboard
5 Gemini 4 Carbon 5.00 Oct 2026 Nov 2026 dubovskylabs.com Official leaderboard
6 Claude Fable 5.1 6.00 one-shot Sep 2026 Sep 2026 dubovskylabs.com Official leaderboard
7 Claude Fable 5 7.00 Jun 2026 dubovskylabs.com Official leaderboard
8 DeepSeek V4.1 Flash 1017 8.00 Oct 2026 Oct 2026 dubovskylabs.com Official leaderboard
9 Kimi K3 9.00 Jul 2026 dubovskylabs.com Official leaderboard
10 GLM-5.3-Flash 10.0 multi-prompt Aug 2026 Aug 2026 dubovskylabs.com Official leaderboard
11 GPT-6 Astra 11.0 one-shot Sep 2026 Sep 2026 dubovskylabs.com Official leaderboard
12 Union Alpha closed weights 12.0 multi-prompt Sep 2026 dubovskylabs.com Official leaderboard
13 Qwen3.8-Flash-Next 13.0 Q2, multi-prompt Aug 2026 Sep 2026 dubovskylabs.com Official leaderboard
14 DeepSeek V4.1 Flash 14.0 one-shot Sep 2026 Sep 2026 dubovskylabs.com Official leaderboard
15 Gemini 4 Argon 15.0 Sep 2026 Oct 2026 dubovskylabs.com Official leaderboard
16 Qwen 3.8 27B 16.0 ThinkingCap, multi-prompt Aug 2026 Sep 2026 dubovskylabs.com Official leaderboard
17 GPT-6 Sol 17.0 Sep 2026 Sep 2026 dubovskylabs.com Official leaderboard
18 GPT-6.1 Sol 18.0 Sep 2026 Sep 2026 dubovskylabs.com Official leaderboard
19 GPT-5.6 Sol 19.0 Jul 2026 dubovskylabs.com Official leaderboard
20 Claude Opus 4.8 20.0 May 2026 dubovskylabs.com Official leaderboard
21 Muse Spark 1.3 21.0 one-shot Sep 2026 Sep 2026 dubovskylabs.com Official leaderboard
22 Grok 4.7 22.0 Sep 2026 Oct 2026 dubovskylabs.com Official leaderboard
23 Grok 4.6 23.0 one-shot Aug 2026 Sep 2026 dubovskylabs.com Official leaderboard
24 Grok 4.5 24.0 Jul 2026 dubovskylabs.com Official leaderboard
25 Gemini 3.8 Flash 25.0 one-shot Sep 2026 Sep 2026 dubovskylabs.com Official leaderboard
26 Gemini 3.5 Flash 26.0 May 2026 dubovskylabs.com Official leaderboard
27 GPT-6 Luna 27.0 multi-prompt Sep 2026 Sep 2026 dubovskylabs.com Official leaderboard
28 Gemini 3.6 Flash 29.0 Jul 2026 dubovskylabs.com Official leaderboard
29 Qwen 3.6 27B ThinkingCap open weights, 27B 30.0 ThinkingCap dubovskylabs.com Official leaderboard
30 Muse Spark 1.2 31.0 Aug 2026 dubovskylabs.com Official leaderboard
31 GPT-5.6 Terra 32.0 one-shot Jul 2026 Aug 2026 dubovskylabs.com Official leaderboard
32 Qwen3.5 397B-A17B 33.0 Feb 2026 dubovskylabs.com Official leaderboard
33 MiniMax-M2.7 34.0 Mar 2026 dubovskylabs.com Official leaderboard
34 Qwen 3.8 Max 35.0 Aug 2026 dubovskylabs.com Official leaderboard
35 DeepSeek-V4-Pro 36.0 0831, multi-prompt Apr 2026 Aug 2026 dubovskylabs.com Official leaderboard
36 DeepSeek V4 Flash 0731 37.0 DS4, one-shot Jul 2026 Aug 2026 dubovskylabs.com Official leaderboard
37 Gemini 3.7 Flash 40.0 multi-prompt Aug 2026 Aug 2026 dubovskylabs.com Official leaderboard
38 DeepSeek-V4-Flash-Vision-Exp 41.0 multi-prompt Aug 2026 Sep 2026 dubovskylabs.com Official leaderboard
39 GPT-5.5 42.0 Apr 2026 dubovskylabs.com Official leaderboard
40 MiniMax-M3 43.0 Jun 2026 dubovskylabs.com Official leaderboard
41 Claude Sonnet 4.6 44.0 Feb 2026 dubovskylabs.com Official leaderboard
42 GLM-5.2 45.0 Jun 2026 dubovskylabs.com Official leaderboard
43 MiMo-V2.6 Pro 46.0 Sep 2026 Sep 2026 dubovskylabs.com Official leaderboard
44 MiMo-V2.6 Flash 47.0 Sep 2026 Sep 2026 dubovskylabs.com Official leaderboard
45 Mistral Large 4 48.0 Oct 2026 Oct 2026 dubovskylabs.com Official leaderboard
46 DeepSeek-V4-Flash 49.0 Apr 2026 dubovskylabs.com Official leaderboard
47 Qwen3.7-Max 50.0 May 2026 dubovskylabs.com Official leaderboard
48 Big Pickle 52.0 dubovskylabs.com Official leaderboard
49 Kimi K2.7 Code 53.0 Jun 2026 dubovskylabs.com Official leaderboard
50 Qwen3.6 27B 54.0 MLX, multi-prompt Apr 2026 Sep 2026 dubovskylabs.com Official leaderboard
51 Qwen3.5-122B-A10B 56.0 Feb 2026 dubovskylabs.com Official leaderboard
52 glm-4.7-flash 57.0 Jan 2026 dubovskylabs.com Official leaderboard
53 Qwen 3.6 35B-A3B 58.0 Apr 2026 dubovskylabs.com Official leaderboard
54 Muse Glimmer 59.0 one-shot Aug 2026 Aug 2026 dubovskylabs.com Official leaderboard
55 Gemma 4 31B IT 60.0 Apr 2026 dubovskylabs.com Official leaderboard
56 Laguna S 2.1 61.0 multi-prompt Jul 2026 Aug 2026 dubovskylabs.com Official leaderboard
57 GPT-5.6 Luna 62.0 one-shot Jul 2026 Aug 2026 dubovskylabs.com Official leaderboard
58 MiMo v5 open weights 63.0 dubovskylabs.com Official leaderboard
59 Mistral Small 4 64.0 Mar 2026 dubovskylabs.com Official leaderboard
60 Ornith 1.0 35B open weights, 35B 65.0 dubovskylabs.com Official leaderboard
61 Qwen3.5-35B-A3B 66.0 Feb 2026 dubovskylabs.com Official leaderboard
62 Nemotron 3 Ultra 68.0 Jun 2026 Official leaderboard
#ModelScoreSettingReleasedRun onLinkSource
1 Claude Sonnet 5.5 1.00 Sep 2026 Oct 2026 dubovskylabs.com Official leaderboard
2 Claude Opus 5.5 2.00 multi-prompt Sep 2026 Sep 2026 dubovskylabs.com Official leaderboard
3 Claude Fable 5.2 3.00 Nov 2026 Nov 2026 dubovskylabs.com Official leaderboard
4 Claude Opus 5 4.00 Jul 2026 dubovskylabs.com Official leaderboard
5 Gemini 4 Carbon 5.00 Oct 2026 Nov 2026 dubovskylabs.com Official leaderboard
6 Claude Fable 5.1 6.00 one-shot Sep 2026 Sep 2026 dubovskylabs.com Official leaderboard
7 Claude Fable 5 7.00 Jun 2026 dubovskylabs.com Official leaderboard
8 DeepSeek V4.1 Flash 1017 8.00 Oct 2026 Oct 2026 dubovskylabs.com Official leaderboard
9 Kimi K3 9.00 Jul 2026 dubovskylabs.com Official leaderboard
10 GLM-5.3-Flash 10.0 multi-prompt Aug 2026 Aug 2026 dubovskylabs.com Official leaderboard
11 GPT-6 Astra 11.0 one-shot Sep 2026 Sep 2026 dubovskylabs.com Official leaderboard
12 Qwen3.8-Flash-Next 13.0 Q2, multi-prompt Aug 2026 Sep 2026 dubovskylabs.com Official leaderboard
13 DeepSeek V4.1 Flash 14.0 one-shot Sep 2026 Sep 2026 dubovskylabs.com Official leaderboard
14 Gemini 4 Argon 15.0 Sep 2026 Oct 2026 dubovskylabs.com Official leaderboard
15 Qwen 3.8 27B 16.0 ThinkingCap, multi-prompt Aug 2026 Sep 2026 dubovskylabs.com Official leaderboard
16 GPT-6 Sol 17.0 Sep 2026 Sep 2026 dubovskylabs.com Official leaderboard
17 GPT-6.1 Sol 18.0 Sep 2026 Sep 2026 dubovskylabs.com Official leaderboard
18 GPT-5.6 Sol 19.0 Jul 2026 dubovskylabs.com Official leaderboard
19 Claude Opus 4.8 20.0 May 2026 dubovskylabs.com Official leaderboard
20 Muse Spark 1.3 21.0 one-shot Sep 2026 Sep 2026 dubovskylabs.com Official leaderboard
21 Grok 4.7 22.0 Sep 2026 Oct 2026 dubovskylabs.com Official leaderboard
22 Grok 4.6 23.0 one-shot Aug 2026 Sep 2026 dubovskylabs.com Official leaderboard
23 Grok 4.5 24.0 Jul 2026 dubovskylabs.com Official leaderboard
24 Gemini 3.8 Flash 25.0 one-shot Sep 2026 Sep 2026 dubovskylabs.com Official leaderboard
25 Gemini 3.5 Flash 26.0 May 2026 dubovskylabs.com Official leaderboard
26 GPT-6 Luna 27.0 multi-prompt Sep 2026 Sep 2026 dubovskylabs.com Official leaderboard
27 Gemini 3.6 Flash 29.0 Jul 2026 dubovskylabs.com Official leaderboard
28 Muse Spark 1.2 31.0 Aug 2026 dubovskylabs.com Official leaderboard
29 GPT-5.6 Terra 32.0 one-shot Jul 2026 Aug 2026 dubovskylabs.com Official leaderboard
30 Qwen3.5 397B-A17B 33.0 Feb 2026 dubovskylabs.com Official leaderboard
31 MiniMax-M2.7 34.0 Mar 2026 dubovskylabs.com Official leaderboard
32 Qwen 3.8 Max 35.0 Aug 2026 dubovskylabs.com Official leaderboard
33 DeepSeek-V4-Pro 36.0 0831, multi-prompt Apr 2026 Aug 2026 dubovskylabs.com Official leaderboard
34 DeepSeek V4 Flash 0731 37.0 DS4, one-shot Jul 2026 Aug 2026 dubovskylabs.com Official leaderboard
35 Gemini 3.7 Flash 40.0 multi-prompt Aug 2026 Aug 2026 dubovskylabs.com Official leaderboard
36 DeepSeek-V4-Flash-Vision-Exp 41.0 multi-prompt Aug 2026 Sep 2026 dubovskylabs.com Official leaderboard
37 GPT-5.5 42.0 Apr 2026 dubovskylabs.com Official leaderboard
38 MiniMax-M3 43.0 Jun 2026 dubovskylabs.com Official leaderboard
39 Claude Sonnet 4.6 44.0 Feb 2026 dubovskylabs.com Official leaderboard
40 GLM-5.2 45.0 Jun 2026 dubovskylabs.com Official leaderboard
41 MiMo-V2.6 Pro 46.0 Sep 2026 Sep 2026 dubovskylabs.com Official leaderboard
42 MiMo-V2.6 Flash 47.0 Sep 2026 Sep 2026 dubovskylabs.com Official leaderboard
43 Mistral Large 4 48.0 Oct 2026 Oct 2026 dubovskylabs.com Official leaderboard
44 DeepSeek-V4-Flash 49.0 Apr 2026 dubovskylabs.com Official leaderboard
45 Qwen3.7-Max 50.0 May 2026 dubovskylabs.com Official leaderboard
46 Kimi K2.7 Code 53.0 Jun 2026 dubovskylabs.com Official leaderboard
47 Qwen3.6 27B 54.0 MLX, multi-prompt Apr 2026 Sep 2026 dubovskylabs.com Official leaderboard
48 Qwen3.5-122B-A10B 56.0 Feb 2026 dubovskylabs.com Official leaderboard
49 glm-4.7-flash 57.0 Jan 2026 dubovskylabs.com Official leaderboard
50 Qwen 3.6 35B-A3B 58.0 Apr 2026 dubovskylabs.com Official leaderboard
51 Muse Glimmer 59.0 one-shot Aug 2026 Aug 2026 dubovskylabs.com Official leaderboard
52 Gemma 4 31B IT 60.0 Apr 2026 dubovskylabs.com Official leaderboard
53 Laguna S 2.1 61.0 multi-prompt Jul 2026 Aug 2026 dubovskylabs.com Official leaderboard
54 GPT-5.6 Luna 62.0 one-shot Jul 2026 Aug 2026 dubovskylabs.com Official leaderboard
55 Mistral Small 4 64.0 Mar 2026 dubovskylabs.com Official leaderboard
56 Qwen3.5-35B-A3B 66.0 Feb 2026 dubovskylabs.com Official leaderboard
57 Nemotron 3 Ultra 68.0 Jun 2026 Official leaderboard

Scores at each effort level

Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.

Lowest effort to highest, light to dark
Qwen3.6 27B
MLX, multi-… GGUF
67.0
DeepSeek-V4-Pro
0831, multi…
51.0
Qwen 3.8 27B
ThinkingCap… multi-prompt
28.0
Claude Fable 5
UltraCode
39.0

Source: Official BAAnchmark leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
Nemotron 3 Ultra 68.0
Qwen3.6 27B GGUF 67.0
Qwen3.5-35B-A3B 66.0
Mistral Small 4 64.0
GPT-5.6 Luna one-shot 62.0
Laguna S 2.1 multi-prompt 61.0
Gemma 4 31B IT 60.0
Muse Glimmer one-shot 59.0
Qwen 3.6 35B-A3B 58.0
glm-4.7-flash 57.0
Qwen3.5-122B-A10B 56.0
Qwen3.6 27B 55.0
Qwen3.6 27B MLX, multi-prompt 54.0
Kimi K2.7 Code 53.0
DeepSeek-V4-Pro 51.0
Qwen3.7-Max 50.0
DeepSeek-V4-Flash 49.0
Mistral Large 4 48.0
MiMo-V2.6 Flash 47.0
MiMo-V2.6 Pro 46.0
GLM-5.2 45.0
Claude Sonnet 4.6 44.0
MiniMax-M3 43.0
GPT-5.5 42.0
DeepSeek-V4-Flash-Vision-Exp multi-prompt 41.0
Gemini 3.7 Flash multi-prompt 40.0
Claude Fable 5 UltraCode 39.0
GLM-5.3-Flash multi-prompt Claude Code 38.0
DeepSeek V4 Flash 0731 DS4, one-shot 37.0
DeepSeek-V4-Pro 0831, multi-prompt 36.0
Qwen 3.8 Max 35.0
MiniMax-M2.7 34.0
Qwen3.5 397B-A17B 33.0
GPT-5.6 Terra one-shot 32.0
Muse Spark 1.2 31.0
Gemini 3.6 Flash 29.0
Qwen 3.8 27B multi-prompt 28.0
GPT-6 Luna multi-prompt 27.0
Gemini 3.5 Flash 26.0
Gemini 3.8 Flash one-shot 25.0
Grok 4.5 24.0
Grok 4.6 one-shot 23.0
Grok 4.7 22.0
Muse Spark 1.3 one-shot 21.0
Claude Opus 4.8 20.0
GPT-5.6 Sol 19.0
GPT-6.1 Sol 18.0
GPT-6 Sol 17.0
Qwen 3.8 27B ThinkingCap, multi-prompt 16.0
Gemini 4 Argon 15.0
DeepSeek V4.1 Flash one-shot 14.0
Qwen3.8-Flash-Next Q2, multi-prompt 13.0
GPT-6 Astra one-shot 11.0
GLM-5.3-Flash multi-prompt 10.0
Kimi K3 9.00
DeepSeek V4.1 Flash 1017 8.00
Claude Fable 5 7.00
Claude Fable 5.1 one-shot 6.00
Gemini 4 Carbon 5.00
Claude Opus 5 4.00
Claude Fable 5.2 3.00
Claude Opus 5.5 multi-prompt 2.00
Claude Sonnet 5.5 1.00

Scores at each effort level

Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.

Lowest effort to highest, light to dark
Qwen3.6 27B
MLX, multi-… GGUF
67.0
DeepSeek-V4-Pro
0831, multi…
51.0
Qwen 3.8 27B
ThinkingCap… multi-prompt
28.0
Claude Fable 5
UltraCode
39.0

Source: Official BAAnchmark leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
Nemotron 3 Ultra 68.0
Qwen3.6 27B GGUF 67.0
Qwen3.5-35B-A3B 66.0
Mistral Small 4 64.0
GPT-5.6 Luna one-shot 62.0
Laguna S 2.1 multi-prompt 61.0
Gemma 4 31B IT 60.0
Muse Glimmer one-shot 59.0
Qwen 3.6 35B-A3B 58.0
glm-4.7-flash 57.0
Qwen3.5-122B-A10B 56.0
Qwen3.6 27B 55.0
Qwen3.6 27B MLX, multi-prompt 54.0
Kimi K2.7 Code 53.0
DeepSeek-V4-Pro 51.0
Qwen3.7-Max 50.0
DeepSeek-V4-Flash 49.0
Mistral Large 4 48.0
MiMo-V2.6 Flash 47.0
MiMo-V2.6 Pro 46.0
GLM-5.2 45.0
Claude Sonnet 4.6 44.0
MiniMax-M3 43.0
GPT-5.5 42.0
DeepSeek-V4-Flash-Vision-Exp multi-prompt 41.0
Gemini 3.7 Flash multi-prompt 40.0
Claude Fable 5 UltraCode 39.0
GLM-5.3-Flash multi-prompt Claude Code 38.0
DeepSeek V4 Flash 0731 DS4, one-shot 37.0
DeepSeek-V4-Pro 0831, multi-prompt 36.0
Qwen 3.8 Max 35.0
MiniMax-M2.7 34.0
Qwen3.5 397B-A17B 33.0
GPT-5.6 Terra one-shot 32.0
Muse Spark 1.2 31.0
Gemini 3.6 Flash 29.0
Qwen 3.8 27B multi-prompt 28.0
GPT-6 Luna multi-prompt 27.0
Gemini 3.5 Flash 26.0
Gemini 3.8 Flash one-shot 25.0
Grok 4.5 24.0
Grok 4.6 one-shot 23.0
Grok 4.7 22.0
Muse Spark 1.3 one-shot 21.0
Claude Opus 4.8 20.0
GPT-5.6 Sol 19.0
GPT-6.1 Sol 18.0
GPT-6 Sol 17.0
Qwen 3.8 27B ThinkingCap, multi-prompt 16.0
Gemini 4 Argon 15.0
DeepSeek V4.1 Flash one-shot 14.0
Qwen3.8-Flash-Next Q2, multi-prompt 13.0
GPT-6 Astra one-shot 11.0
GLM-5.3-Flash multi-prompt 10.0
Kimi K3 9.00
DeepSeek V4.1 Flash 1017 8.00
Claude Fable 5 7.00
Claude Fable 5.1 one-shot 6.00
Gemini 4 Carbon 5.00
Claude Opus 5 4.00
Claude Fable 5.2 3.00
Claude Opus 5.5 multi-prompt 2.00
Claude Sonnet 5.5 1.00

8 models have held number one on the official leaderboard since 2023

Lab claims count from the day the card came out, independent runs from the model’s release.

Z.ai Qwen Google DeepMind Anthropic
Own leaderboard 8 leaders

Source: Official BAAnchmark leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
On the benchmark’s own leaderboard
0.00 20.0 40.0 60.0 80.0 2026 2027 0.00 20.0 40.0 60.0 80.0 2026 2027

Source: Official BAAnchmark leaderboard

No lab model card reports it yet

We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.

More in Coding

All 523 in Coding
  • SWE-bench Verified In use 500 human-validated GitHub issues from 12 Python repos. Agent writes a patch that must pass hidden tests. Claude Opus 4.7 83.5%
  • LiveCodeBench v6 In use Competitive programming problems collected after model cutoffs from LeetCode, AtCoder, Codeforces. Scored pass@1 on hidden tests. DeepSeek-V4-Pro 93.5%
  • SWE-bench Multilingual In use 300 GitHub-issue tasks across nine programming languages in SWE-bench format. Patch graded by repository tests. Claude Fable 5.2 96.6%
  • DeepSWE In use 113 original long-horizon engineering tasks across 91 repos and five languages, written from scratch. Hand-written verifiers accept alternative solutions. GPT-6 Astra 74.1%
  • SciCode In use Scientist-curated research coding problems split into subproblems across natural sciences. Code checked against test cases. Seed2.1 Pro 59.8%
  • Codeforces rating In use Model solves recent Codeforces contest problems. An Elo-style rating is computed from simulated contest performance. DeepSeek V4.1 Flash 3,471

Data audit

All 10 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 4 pass 1 note
  • Pass Independent runs Scores from the official leaderboard (62 entries).
  • Pass Models scored 57 models we track, plus 5 other entries on its leaderboard.
  • Pass Newest score The newest is from November 2026, for Claude Fable 5.2.
  • Pass Frontier models 8 of the top 10 have a score, led by Claude Fable 5.2 (#2).
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on dubovskylabs.com: 63 rows for models we track and 5 other entries at the last read.
  • Pass Last read November 20th, 2026.
  • Pass Read worked The last read worked.
Consistency 1 pass 1 note
  • Note Lab claims No model has both a lab number and an independent run.
  • Pass Known issues No published audit found problems.

How we check