Benchmark Almanac

Game Arena Lines of Action

Gemini 4 Carbon 1,274 leads since October 2026
Not in model cards

Language models play Lines of Action, where a piece moves as far as the number of pieces on its line. The goal is to join all of one's pieces into one group.

Kaggle published the leaderboard on July 16th, 2026 with 36 models on November 1st, 2026. Gameplay logs are a public Kaggle dataset. Part of the Kaggle Game Arena suite.

Made by
Kaggle and Google DeepMind
Released
July 2026
Kind
Puzzles and chess
Level
Everyday
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

Gemini 4 Carbon leads with 1,274

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. 1,274 Gemini 4 Carbon
  2. 1,117 Claude Fable 5.2
  3. 1,067 Gemini 4 Argon
  4. 1,014 GPT-6 Astra
  5. 866 GPT-6.1 Sol
  6. 822 Claude Opus 5
  7. 655 Claude Fable 5.1
  8. 607 Claude Opus 5.5
  9. 600 GPT-5.5
  10. 570 Grok 4.7
  11. 522 GPT-5.6 Sol
  12. 521 Grok 4.6
  13. 489 GPT-5.4
  14. 469 Grok 4.5
  15. 443 Gemini 3.1 Pro
  16. 435 GPT-6 Sol
  17. 431 Gemini 3.8 Flash
  18. 427 GPT-5.2
  19. 405 Claude Opus 4.8
  20. 393 GPT-6.1 Terra
  21. 380 Claude Opus 4.6
  22. 351 Gemini 3.7 Flash
  23. 350 Gemini 3.6 Flash
  24. 338 Claude Sonnet 5.5
  25. 336 Gemini 3.5 Flash
  26. 321 GPT-5.6 Luna
  27. 310 GPT-5.6 Terra
  28. 293 o3
  29. 284 Grok 4.20
  30. 274 GPT-5.4 Mini
  31. 226 Claude Sonnet 4.5
  32. 222 Claude Sonnet 5
  33. 209 Claude Sonnet 4.6
  34. 196 GPT-6 Luna
  35. 184 Grok 4.3
  36. 159 Gemini 3 Flash
  37. 138 GPT-5 mini
  38. 45 Claude Haiku 5.5
  39. 0 Claude Opus 4.7
  1. 1 Gemini 4 Carbon 1,274
  2. 2 Claude Fable 5.2 1,117
  3. 3 Gemini 4 Argon 1,067
  4. 4 GPT-6 Astra 1,014
  5. 5 GPT-6.1 Sol 866
  6. 6 Claude Opus 5 822
  7. 7 Claude Fable 5.1 655
  8. 8 Claude Opus 5.5 607
  9. 9 GPT-5.5 600
  10. 10 Grok 4.7 570
  11. 11 GPT-5.6 Sol 522
  12. 12 Grok 4.6 521
  13. 13 GPT-5.4 489
  14. 14 Grok 4.5 469
  15. 15 Gemini 3.1 Pro 443
  16. 16 GPT-6 Sol 435
  17. 17 Gemini 3.8 Flash 431
  18. 18 GPT-5.2 427
  19. 19 Claude Opus 4.8 405
  20. 20 GPT-6.1 Terra 393
  21. 21 Claude Opus 4.6 380
  22. 22 Gemini 3.7 Flash 351
  23. 23 Gemini 3.6 Flash 350
  24. 24 Claude Sonnet 5.5 338
  25. 25 Gemini 3.5 Flash 336
  26. 26 GPT-5.6 Luna 321
  27. 27 GPT-5.6 Terra 310
  28. 28 o3 293
  29. 29 Grok 4.20 reasoning 284
  30. 30 GPT-5.4 Mini 274
  31. 31 Claude Sonnet 4.5 226
  32. 32 Claude Sonnet 5 222
  33. 33 Claude Sonnet 4.6 209
  34. 34 GPT-6 Luna 196
  35. 35 Grok 4.3 184
  36. 36 Gemini 3 Flash 159
  37. 37 GPT-5 mini 138
  38. 38 Claude Haiku 5.5 45
  39. 39 Claude Opus 4.7 0

Source: Official Game Arena Lines of Action leaderboard

#ModelScoreSettingDateSource
1 Gemini 4 Carbon 1,274 Oct 2026 Official leaderboard
2 Claude Fable 5.2 1,117 Nov 2026 Official leaderboard
3 Gemini 4 Argon 1,067 Sep 2026 Official leaderboard
4 GPT-6 Astra 1,014 Sep 2026 Official leaderboard
5 GPT-6.1 Sol 866 Sep 2026 Official leaderboard
6 Claude Opus 5 822 Jul 2026 Official leaderboard
7 Claude Fable 5.1 655 Sep 2026 Official leaderboard
8 Claude Opus 5.5 607 Sep 2026 Official leaderboard
9 GPT-5.5 600 Apr 2026 Official leaderboard
10 Grok 4.7 570 Sep 2026 Official leaderboard
11 GPT-5.6 Sol 522 Jul 2026 Official leaderboard
12 Grok 4.6 521 Aug 2026 Official leaderboard
13 GPT-5.4 489 Mar 2026 Official leaderboard
14 Grok 4.5 469 Jul 2026 Official leaderboard
15 Gemini 3.1 Pro 443 Feb 2026 Official leaderboard
16 GPT-6 Sol 435 Sep 2026 Official leaderboard
17 Gemini 3.8 Flash 431 Sep 2026 Official leaderboard
18 GPT-5.2 427 Dec 2025 Official leaderboard
19 Claude Opus 4.8 405 May 2026 Official leaderboard
20 GPT-6.1 Terra 393 Oct 2026 Official leaderboard
21 Claude Opus 4.6 380 Feb 2026 Official leaderboard
22 Gemini 3.7 Flash 351 Aug 2026 Official leaderboard
23 Gemini 3.6 Flash 350 Jul 2026 Official leaderboard
24 Claude Sonnet 5.5 338 Sep 2026 Official leaderboard
25 Gemini 3.5 Flash 336 May 2026 Official leaderboard
26 GPT-5.6 Luna 321 Jul 2026 Official leaderboard
27 GPT-5.6 Terra 310 Jul 2026 Official leaderboard
28 o3 293 Apr 2025 Official leaderboard
29 Grok 4.20 284 reasoning Feb 2026 Official leaderboard
30 GPT-5.4 Mini 274 Mar 2026 Official leaderboard
31 Claude Sonnet 4.5 226 Sep 2025 Official leaderboard
32 Claude Sonnet 5 222 Jun 2026 Official leaderboard
33 Claude Sonnet 4.6 209 Feb 2026 Official leaderboard
34 GPT-6 Luna 196 Sep 2026 Official leaderboard
35 Grok 4.3 184 Apr 2026 Official leaderboard
36 Gemini 3 Flash 159 Dec 2025 Official leaderboard
37 GPT-5 mini 138 Aug 2025 Official leaderboard
38 Claude Haiku 5.5 45 Oct 2026 Official leaderboard
39 Claude Opus 4.7 0 Apr 2026 Official leaderboard

5 models have held number one on the official leaderboard since its release in July 2026

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

OpenAI Anthropic Google DeepMind
Own leaderboard 5 leaders
Gemini 4 Carbon 1,274

Source: Official Game Arena Lines of Action 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 2025 2026 2027 0 500 1,000 1,500 2025 2026 2027

Source: Official Game Arena Lines of Action 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%