Benchmark Almanac

Game Arena Checkers

GPT-6 Astra 1,338 leads since Sep 2026 Not in model cards

Language models play American checkers against each other, with forced jumps and crowned kings. Ratings come from head-to-head games.

Kaggle published the leaderboard on 2026-06-11 with 38 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 Jun 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
38
Models scored on its own leaderboard

GPT-6 Astra leads with 1,338

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.

38 models. Scroll sideways for the rest.

  1. 1,338 GPT-6 Astra
  2. 1,233 Gemini 3.8 Flash
  3. 1,212 GPT-6.1 Sol
  4. 1,173 Claude Opus 5
  5. 1,130 GPT-5.5
  6. 1,115 Claude Opus 5.5
  7. 1,094 GPT-5.4
  8. 1,080 Claude Opus 4.8
  9. 1,079 Claude Fable 5.1
  10. 1,073 Gemini 3.7 Flash
  11. 1,069 Gemini 4 Argon
  12. 989 o3
  13. 972 Gemini 3.1 Pro
  14. 969 Claude Sonnet 5.5
  15. 967 GPT-5.2
  16. 951 Gemini 3.6 Flash
  17. 949 Claude Sonnet 5
  18. 943 Claude Opus 4.7
  19. 932 GPT-6 Sol
  20. 920 Gemini 3.5 Flash
  21. 892 Gemini 3 Flash
  22. 885 Claude Opus 4.6
  23. 881 GPT-5.6 Sol
  24. 879 Grok 4.7
  25. 744 GPT-5.6 Terra
  26. 690 Claude Sonnet 4.6
  27. 681 GPT-5 mini
  28. 643 Claude Haiku 5.5
  29. 641 GPT-6 Luna
  30. 565 Claude Sonnet 4.5
  31. 553 Grok 4.6
  32. 495 Grok 4.1 Fast
  33. 416 Grok 4.20
  34. 323 GPT-5.6 Luna
  35. 279 GPT-5.4 Mini
  36. 83 Grok 4.5
  37. 59 DeepSeek-V3.2
  38. 0 Grok 4.3
  1. 1 GPT-6 Astra 1,338
  2. 2 Gemini 3.8 Flash 1,233
  3. 3 GPT-6.1 Sol 1,212
  4. 4 Claude Opus 5 1,173
  5. 5 GPT-5.5 1,130
  6. 6 Claude Opus 5.5 1,115
  7. 7 GPT-5.4 1,094
  8. 8 Claude Opus 4.8 1,080
  9. 9 Claude Fable 5.1 1,079
  10. 10 Gemini 3.7 Flash 1,073
  11. 11 Gemini 4 Argon 1,069
  12. 12 o3 989
  13. 13 Gemini 3.1 Pro 972
  14. 14 Claude Sonnet 5.5 969
  15. 15 GPT-5.2 967
  16. 16 Gemini 3.6 Flash 951
  17. 17 Claude Sonnet 5 949
  18. 18 Claude Opus 4.7 943
  19. 19 GPT-6 Sol 932
  20. 20 Gemini 3.5 Flash 920
  21. 21 Gemini 3 Flash 892
  22. 22 Claude Opus 4.6 885
  23. 23 GPT-5.6 Sol 881
  24. 24 Grok 4.7 879
  25. 25 GPT-5.6 Terra 744
  26. 26 Claude Sonnet 4.6 690
  27. 27 GPT-5 mini 681
  28. 28 Claude Haiku 5.5 643
  29. 29 GPT-6 Luna 641
  30. 30 Claude Sonnet 4.5 565
  31. 31 Grok 4.6 553
  32. 32 Grok 4.1 Fast reasoning 495
  33. 33 Grok 4.20 reasoning 416
  34. 34 GPT-5.6 Luna 323
  35. 35 GPT-5.4 Mini 279
  36. 36 Grok 4.5 83
  37. 37 DeepSeek-V3.2 59
  38. 38 Grok 4.3 0
#ModelScoreSettingDateSource
1 GPT-6 Astra 1,338 Sep 2026 Official leaderboard
2 Gemini 3.8 Flash 1,233 Sep 2026 Official leaderboard
3 GPT-6.1 Sol 1,212 Sep 2026 Official leaderboard
4 Claude Opus 5 1,173 Jul 2026 Official leaderboard
5 GPT-5.5 1,130 Apr 2026 Official leaderboard
6 Claude Opus 5.5 1,115 Sep 2026 Official leaderboard
7 GPT-5.4 1,094 Mar 2026 Official leaderboard
8 Claude Opus 4.8 1,080 May 2026 Official leaderboard
9 Claude Fable 5.1 1,079 Sep 2026 Official leaderboard
10 Gemini 3.7 Flash 1,073 Aug 2026 Official leaderboard
11 Gemini 4 Argon 1,069 Sep 2026 Official leaderboard
12 o3 989 Apr 2025 Official leaderboard
13 Gemini 3.1 Pro 972 Feb 2026 Official leaderboard
14 Claude Sonnet 5.5 969 Sep 2026 Official leaderboard
15 GPT-5.2 967 Dec 2025 Official leaderboard
16 Gemini 3.6 Flash 951 Jul 2026 Official leaderboard
17 Claude Sonnet 5 949 Jun 2026 Official leaderboard
18 Claude Opus 4.7 943 Apr 2026 Official leaderboard
19 GPT-6 Sol 932 Sep 2026 Official leaderboard
20 Gemini 3.5 Flash 920 May 2026 Official leaderboard
21 Gemini 3 Flash 892 Dec 2025 Official leaderboard
22 Claude Opus 4.6 885 Feb 2026 Official leaderboard
23 GPT-5.6 Sol 881 Jul 2026 Official leaderboard
24 Grok 4.7 879 Sep 2026 Official leaderboard
25 GPT-5.6 Terra 744 Jul 2026 Official leaderboard
26 Claude Sonnet 4.6 690 Feb 2026 Official leaderboard
27 GPT-5 mini 681 Aug 2025 Official leaderboard
28 Claude Haiku 5.5 643 Oct 2026 Official leaderboard
29 GPT-6 Luna 641 Sep 2026 Official leaderboard
30 Claude Sonnet 4.5 565 Sep 2025 Official leaderboard
31 Grok 4.6 553 Aug 2026 Official leaderboard
32 Grok 4.1 Fast 495 reasoning Nov 2025 Official leaderboard
33 Grok 4.20 416 reasoning Feb 2026 Official leaderboard
34 GPT-5.6 Luna 323 Jul 2026 Official leaderboard
35 GPT-5.4 Mini 279 Mar 2026 Official leaderboard
36 Grok 4.5 83 Jul 2026 Official leaderboard
37 DeepSeek-V3.2 59 Dec 2025 Official leaderboard
38 Grok 4.3 0 Apr 2026 Official leaderboard

4 models have held number one since its release in Jun 2026

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

OpenAI Anthropic Google DeepMind
Own leaderboard 4 leaders
GPT-6 Astra 1,338

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 2025 2026 0 500 1,000 1,500 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