Benchmark Almanac

Game Arena Nine Men's Morris

Claude Opus 5 1,022 leads since Aug 2026 Not in model cards

Language models play Nine Men's Morris. Forming a row of three, called a mill, lets a player remove one of the opponent's pieces.

Kaggle published the leaderboard on 2026-08-04 with 36 models on 2026-11-01. Gameplay logs are a public Kaggle dataset. Part of the Kaggle Game Arena suite.

Made by Kaggle and Google DeepMind Released Aug 2026 Puzzles and chess All tasks public Last model added Oct 2026 Scores refreshed 2 Nov 2026
Level
Everyday Most adults, without training
Scored
Ranked by matches Models play against each other, and the results become ratings.
Answers
Agent task Many steps in a terminal, browser, desktop, app, game or robot.
Kaggle/kaggle-environments 459 stars Dataset Leaderboard kaggle.com
36
Models scored on its own leaderboard

Claude Opus 5 leads with 1,022

Best score per model from the benchmark's own leaderboard, read on 2 Nov 2026. Where a model ran with several settings or harnesses, we keep its best.

36 models. Scroll sideways for the rest.

  1. 1,022 Claude Opus 5
  2. 1,017 Claude Fable 5.1
  3. 989 Gemini 4 Argon
  4. 975 Claude Opus 5.5
  5. 927 Claude Sonnet 5.5
  6. 908 Grok 4.7
  7. 817 Claude Opus 4.8
  8. 788 Gemini 3.1 Pro
  9. 787 Claude Haiku 5.5
  10. 779 Gemini 3.8 Flash
  11. 766 GPT-6.1 Sol
  12. 762 GPT-6 Astra
  13. 745 Claude Sonnet 5
  14. 705 Gemini 3.7 Flash
  15. 678 Grok 4.6
  16. 637 GPT-5.6 Sol
  17. 620 GPT-5.5
  18. 607 GPT-5.4
  19. 594 Gemini 3.6 Flash
  20. 582 Gemini 3 Flash
  21. 568 GPT-6 Sol
  22. 561 o3
  23. 542 Claude Opus 4.7
  24. 527 Grok 4.5
  25. 523 Claude Opus 4.6
  26. 520 Gemini 3.5 Flash
  27. 510 GPT-5.2
  28. 426 Claude Sonnet 4.6
  29. 375 Grok 4.20
  30. 356 Claude Sonnet 4.5
  31. 289 GPT-5 mini
  32. 260 GPT-5.6 Terra
  33. 208 GPT-5.4 Mini
  34. 33 GPT-5.6 Luna
  35. 12 Grok 4.3
  36. 0 GPT-6 Luna
  1. 1 Claude Opus 5 1,022
  2. 2 Claude Fable 5.1 1,017
  3. 3 Gemini 4 Argon 989
  4. 4 Claude Opus 5.5 975
  5. 5 Claude Sonnet 5.5 927
  6. 6 Grok 4.7 908
  7. 7 Claude Opus 4.8 817
  8. 8 Gemini 3.1 Pro 788
  9. 9 Claude Haiku 5.5 787
  10. 10 Gemini 3.8 Flash 779
  11. 11 GPT-6.1 Sol 766
  12. 12 GPT-6 Astra 762
  13. 13 Claude Sonnet 5 745
  14. 14 Gemini 3.7 Flash 705
  15. 15 Grok 4.6 678
  16. 16 GPT-5.6 Sol 637
  17. 17 GPT-5.5 620
  18. 18 GPT-5.4 607
  19. 19 Gemini 3.6 Flash 594
  20. 20 Gemini 3 Flash 582
  21. 21 GPT-6 Sol 568
  22. 22 o3 561
  23. 23 Claude Opus 4.7 542
  24. 24 Grok 4.5 527
  25. 25 Claude Opus 4.6 523
  26. 26 Gemini 3.5 Flash 520
  27. 27 GPT-5.2 510
  28. 28 Claude Sonnet 4.6 426
  29. 29 Grok 4.20 reasoning 375
  30. 30 Claude Sonnet 4.5 356
  31. 31 GPT-5 mini 289
  32. 32 GPT-5.6 Terra 260
  33. 33 GPT-5.4 Mini 208
  34. 34 GPT-5.6 Luna 33
  35. 35 Grok 4.3 12
  36. 36 GPT-6 Luna 0
#ModelScoreSettingDateSource
1 Claude Opus 5 1,022 Jul 2026 Official leaderboard
2 Claude Fable 5.1 1,017 Sep 2026 Official leaderboard
3 Gemini 4 Argon 989 Sep 2026 Official leaderboard
4 Claude Opus 5.5 975 Sep 2026 Official leaderboard
5 Claude Sonnet 5.5 927 Sep 2026 Official leaderboard
6 Grok 4.7 908 Sep 2026 Official leaderboard
7 Claude Opus 4.8 817 May 2026 Official leaderboard
8 Gemini 3.1 Pro 788 Feb 2026 Official leaderboard
9 Claude Haiku 5.5 787 Oct 2026 Official leaderboard
10 Gemini 3.8 Flash 779 Sep 2026 Official leaderboard
11 GPT-6.1 Sol 766 Sep 2026 Official leaderboard
12 GPT-6 Astra 762 Sep 2026 Official leaderboard
13 Claude Sonnet 5 745 Jun 2026 Official leaderboard
14 Gemini 3.7 Flash 705 Aug 2026 Official leaderboard
15 Grok 4.6 678 Aug 2026 Official leaderboard
16 GPT-5.6 Sol 637 Jul 2026 Official leaderboard
17 GPT-5.5 620 Apr 2026 Official leaderboard
18 GPT-5.4 607 Mar 2026 Official leaderboard
19 Gemini 3.6 Flash 594 Jul 2026 Official leaderboard
20 Gemini 3 Flash 582 Dec 2025 Official leaderboard
21 GPT-6 Sol 568 Sep 2026 Official leaderboard
22 o3 561 Apr 2025 Official leaderboard
23 Claude Opus 4.7 542 Apr 2026 Official leaderboard
24 Grok 4.5 527 Jul 2026 Official leaderboard
25 Claude Opus 4.6 523 Feb 2026 Official leaderboard
26 Gemini 3.5 Flash 520 May 2026 Official leaderboard
27 GPT-5.2 510 Dec 2025 Official leaderboard
28 Claude Sonnet 4.6 426 Feb 2026 Official leaderboard
29 Grok 4.20 375 reasoning Feb 2026 Official leaderboard
30 Claude Sonnet 4.5 356 Sep 2025 Official leaderboard
31 GPT-5 mini 289 Aug 2025 Official leaderboard
32 GPT-5.6 Terra 260 Jul 2026 Official leaderboard
33 GPT-5.4 Mini 208 Mar 2026 Official leaderboard
34 GPT-5.6 Luna 33 Jul 2026 Official leaderboard
35 Grok 4.3 12 Apr 2026 Official leaderboard
36 GPT-6 Luna 0 Sep 2026 Official leaderboard

Number one over time

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

Anthropic
Own leaderboard 1 leader
Claude Opus 5 1,022

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
0 250 500 750 1,000 1,250 2025 2026 0 250 500 750 1,000 1,250 2025 2026

No lab model card reports it yet

We read the headline results of 479 launch posts from 51 labs, and none of them list this benchmark.

More in Games and long tasks

All 68 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 100%
  • 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%
  • Orak Fading LLM agents play 12 popular video games across most genres through an MCP interface. Scores come from in-game progress or outcomes per title. Nemotron 3 Nano 4B 22.9
  • GameWorld Only in system cards Multimodal game agents play 34 browser games across 170 tasks. Outcomes are checked from the game state. Gemini 3 Flash 41.9%
  • 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%
  • TextQuests Only in system cards Agents play 25 classic interactive fiction games. Measures game progress and harmful actions per 100 steps. Muse Spark 2.40