Benchmark Almanac

Game Arena Go

Claude Fable 5.2 1,877 leads since November 2026
Not in model cards

Language models play Go against each other through the Game Arena text harness. Ratings come from head-to-head games.

Google promised Go at the 2025-08 launch. The leaderboard was published on September 22nd, 2026 with 36 models.

Made by
Kaggle and Google DeepMind
Released
September 2026
Kind
Puzzles and chess
Level
Expert
Scored
Ranked by matches
Answers
Agent task
Tasks
All public
Strongest model tested
Claude Fable 5.2 #1 on the capability index
Last model added
November 2026
Scores refreshed
November 12th, 2026
39
Models scored on its own leaderboard

Claude Fable 5.2 leads with 1,877

Best score per model from the benchmark's own leaderboard, read on November 12th, 2026. Where a model ran with several settings or harnesses, we keep its best.

39 models. Scroll sideways for the rest.

  1. 1,877 Claude Fable 5.2
  2. 1,830 Gemini 4 Carbon
  3. 1,735 GPT-6 Astra
  4. 1,686 Gemini 4 Argon
  5. 1,601 Claude Fable 5.1
  6. 1,564 Claude Opus 5
  7. 1,543 Claude Opus 5.5
  8. 1,448 GPT-6.1 Sol
  9. 1,442 Claude Sonnet 5.5
  10. 1,202 GPT-6.1 Terra
  11. 1,170 Gemini 3.8 Flash
  12. 1,147 Gemini 3.1 Pro
  13. 1,118 Claude Sonnet 5
  14. 1,116 Claude Opus 4.8
  15. 1,076 Claude Haiku 5.5
  16. 1,023 Grok 4.6
  17. 1,006 Gemini 3.7 Flash
  18. 1,005 GPT-5.6 Sol
  19. 1,004 Grok 4.7
  20. 996 GPT-5.5
  21. 904 Grok 4.5
  22. 900 GPT-6 Sol
  23. 856 GPT-5.4
  24. 756 Gemini 3.5 Flash
  25. 712 Gemini 3.6 Flash
  26. 570 Claude Opus 4.6
  27. 561 GPT-5.2
  28. 555 o3
  29. 512 GPT-5.6 Terra
  30. 510 Claude Opus 4.7
  31. 497 Gemini 3 Flash
  32. 474 GPT-6 Luna
  33. 471 Grok 4.20
  34. 446 GPT-5.6 Luna
  35. 325 Claude Sonnet 4.5
  36. 296 Claude Sonnet 4.6
  37. 130 Grok 4.3
  38. 120 GPT-5.4 Mini
  39. 0 GPT-5 mini
  1. 1 Claude Fable 5.2 1,877
  2. 2 Gemini 4 Carbon 1,830
  3. 3 GPT-6 Astra 1,735
  4. 4 Gemini 4 Argon 1,686
  5. 5 Claude Fable 5.1 1,601
  6. 6 Claude Opus 5 1,564
  7. 7 Claude Opus 5.5 1,543
  8. 8 GPT-6.1 Sol 1,448
  9. 9 Claude Sonnet 5.5 1,442
  10. 10 GPT-6.1 Terra 1,202
  11. 11 Gemini 3.8 Flash 1,170
  12. 12 Gemini 3.1 Pro 1,147
  13. 13 Claude Sonnet 5 1,118
  14. 14 Claude Opus 4.8 1,116
  15. 15 Claude Haiku 5.5 1,076
  16. 16 Grok 4.6 1,023
  17. 17 Gemini 3.7 Flash 1,006
  18. 18 GPT-5.6 Sol 1,005
  19. 19 Grok 4.7 1,004
  20. 20 GPT-5.5 996
  21. 21 Grok 4.5 904
  22. 22 GPT-6 Sol 900
  23. 23 GPT-5.4 856
  24. 24 Gemini 3.5 Flash 756
  25. 25 Gemini 3.6 Flash 712
  26. 26 Claude Opus 4.6 570
  27. 27 GPT-5.2 561
  28. 28 o3 555
  29. 29 GPT-5.6 Terra 512
  30. 30 Claude Opus 4.7 510
  31. 31 Gemini 3 Flash 497
  32. 32 GPT-6 Luna 474
  33. 33 Grok 4.20 reasoning 471
  34. 34 GPT-5.6 Luna 446
  35. 35 Claude Sonnet 4.5 325
  36. 36 Claude Sonnet 4.6 296
  37. 37 Grok 4.3 130
  38. 38 GPT-5.4 Mini 120
  39. 39 GPT-5 mini 0

Source: Official Game Arena Go leaderboard

#ModelScoreSettingDateSource
1 Claude Fable 5.2 1,877 Nov 2026 Official leaderboard
2 Gemini 4 Carbon 1,830 Oct 2026 Official leaderboard
3 GPT-6 Astra 1,735 Sep 2026 Official leaderboard
4 Gemini 4 Argon 1,686 Sep 2026 Official leaderboard
5 Claude Fable 5.1 1,601 Sep 2026 Official leaderboard
6 Claude Opus 5 1,564 Jul 2026 Official leaderboard
7 Claude Opus 5.5 1,543 Sep 2026 Official leaderboard
8 GPT-6.1 Sol 1,448 Sep 2026 Official leaderboard
9 Claude Sonnet 5.5 1,442 Sep 2026 Official leaderboard
10 GPT-6.1 Terra 1,202 Oct 2026 Official leaderboard
11 Gemini 3.8 Flash 1,170 Sep 2026 Official leaderboard
12 Gemini 3.1 Pro 1,147 Feb 2026 Official leaderboard
13 Claude Sonnet 5 1,118 Jun 2026 Official leaderboard
14 Claude Opus 4.8 1,116 May 2026 Official leaderboard
15 Claude Haiku 5.5 1,076 Oct 2026 Official leaderboard
16 Grok 4.6 1,023 Aug 2026 Official leaderboard
17 Gemini 3.7 Flash 1,006 Aug 2026 Official leaderboard
18 GPT-5.6 Sol 1,005 Jul 2026 Official leaderboard
19 Grok 4.7 1,004 Sep 2026 Official leaderboard
20 GPT-5.5 996 Apr 2026 Official leaderboard
21 Grok 4.5 904 Jul 2026 Official leaderboard
22 GPT-6 Sol 900 Sep 2026 Official leaderboard
23 GPT-5.4 856 Mar 2026 Official leaderboard
24 Gemini 3.5 Flash 756 May 2026 Official leaderboard
25 Gemini 3.6 Flash 712 Jul 2026 Official leaderboard
26 Claude Opus 4.6 570 Feb 2026 Official leaderboard
27 GPT-5.2 561 Dec 2025 Official leaderboard
28 o3 555 Apr 2025 Official leaderboard
29 GPT-5.6 Terra 512 Jul 2026 Official leaderboard
30 Claude Opus 4.7 510 Apr 2026 Official leaderboard
31 Gemini 3 Flash 497 Dec 2025 Official leaderboard
32 GPT-6 Luna 474 Sep 2026 Official leaderboard
33 Grok 4.20 471 reasoning Feb 2026 Official leaderboard
34 GPT-5.6 Luna 446 Jul 2026 Official leaderboard
35 Claude Sonnet 4.5 325 Sep 2025 Official leaderboard
36 Claude Sonnet 4.6 296 Feb 2026 Official leaderboard
37 Grok 4.3 130 Apr 2026 Official leaderboard
38 GPT-5.4 Mini 120 Mar 2026 Official leaderboard
39 GPT-5 mini 0 Aug 2025 Official leaderboard

4 models have held number one on the official leaderboard since its release in September 2026

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

Anthropic OpenAI Google DeepMind
Own leaderboard 4 leaders
Claude Fable 5.2 1,877

Source: Official Game Arena Go leaderboard

Scores over time

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

On the benchmark’s own leaderboard
0 500 1,000 1,500 2,000 2025 2026 2027 0 500 1,000 1,500 2,000 2025 2026 2027

Source: Official Game Arena Go leaderboard

No lab model card reports it yet

We read the headline results of 522 launch posts from 52 labs, and none of them list this benchmark.

More in Games and long tasks

All 82 in Games and long tasks
  • ARC-AGI-3 In use Interactive game environments where agents must explore, infer goals and rules, and act efficiently without instructions. GPT-6 Astra 99.95%
  • Orak Fading LLM agents play 12 popular video games across most genres through an MCP interface. Each game scores in-game progress or outcomes, and the board ranks models by their average rank across the games. Gemini 2.5 Pro 3.90
  • MineDojo Fading Agents complete open-ended tasks in Minecraft, such as mining blocks and fighting mobs, from a large simulation suite. Programmatic checks score many tasks. Seed2.0 Pro 49%
  • TextQuests Only in system cards Agents play 25 classic interactive fiction games. The score is harmful actions per 100 steps, where lower is better. The board also reports game progress. Llama 4 Scout 7.00
  • GameWorld Only in system cards Multimodal game agents play 34 browser games across 170 tasks. Outcomes are checked from the game state. GPT-6.1 Terra 76.8%
  • Swarm Bench Only in system cards Models orchestrate agent swarms on complex tasks that benefit from coordinated decomposition and parallel execution. Kimi K3 76.3%