Benchmark Almanac

Game Arena Lines of Action

Gemini 4 Argon 1,185 leads since Sep 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 2026-07-16 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 Jul 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

Gemini 4 Argon leads with 1,185

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. 1,185 Gemini 4 Argon
  2. 1,127 GPT-6 Astra
  3. 973 GPT-6.1 Sol
  4. 938 Claude Opus 5
  5. 774 Claude Fable 5.1
  6. 723 Claude Opus 5.5
  7. 714 GPT-5.5
  8. 691 Grok 4.7
  9. 657 Grok 4.6
  10. 638 GPT-5.6 Sol
  11. 609 GPT-5.4
  12. 585 Grok 4.5
  13. 570 Gemini 3.1 Pro
  14. 551 GPT-6 Sol
  15. 549 Gemini 3.8 Flash
  16. 543 GPT-5.2
  17. 522 Claude Opus 4.8
  18. 496 Claude Opus 4.6
  19. 480 Gemini 3.6 Flash
  20. 471 Gemini 3.7 Flash
  21. 460 Claude Sonnet 5.5
  22. 456 Gemini 3.5 Flash
  23. 439 GPT-5.6 Luna
  24. 433 GPT-5.6 Terra
  25. 412 o3
  26. 403 Grok 4.20
  27. 392 GPT-5.4 Mini
  28. 344 Claude Sonnet 4.5
  29. 339 Claude Sonnet 5
  30. 323 Claude Sonnet 4.6
  31. 319 GPT-6 Luna
  32. 293 Grok 4.3
  33. 277 Gemini 3 Flash
  34. 254 GPT-5 mini
  35. 177 Claude Haiku 5.5
  36. 0 Claude Opus 4.7
  1. 1 Gemini 4 Argon 1,185
  2. 2 GPT-6 Astra 1,127
  3. 3 GPT-6.1 Sol 973
  4. 4 Claude Opus 5 938
  5. 5 Claude Fable 5.1 774
  6. 6 Claude Opus 5.5 723
  7. 7 GPT-5.5 714
  8. 8 Grok 4.7 691
  9. 9 Grok 4.6 657
  10. 10 GPT-5.6 Sol 638
  11. 11 GPT-5.4 609
  12. 12 Grok 4.5 585
  13. 13 Gemini 3.1 Pro 570
  14. 14 GPT-6 Sol 551
  15. 15 Gemini 3.8 Flash 549
  16. 16 GPT-5.2 543
  17. 17 Claude Opus 4.8 522
  18. 18 Claude Opus 4.6 496
  19. 19 Gemini 3.6 Flash 480
  20. 20 Gemini 3.7 Flash 471
  21. 21 Claude Sonnet 5.5 460
  22. 22 Gemini 3.5 Flash 456
  23. 23 GPT-5.6 Luna 439
  24. 24 GPT-5.6 Terra 433
  25. 25 o3 412
  26. 26 Grok 4.20 reasoning 403
  27. 27 GPT-5.4 Mini 392
  28. 28 Claude Sonnet 4.5 344
  29. 29 Claude Sonnet 5 339
  30. 30 Claude Sonnet 4.6 323
  31. 31 GPT-6 Luna 319
  32. 32 Grok 4.3 293
  33. 33 Gemini 3 Flash 277
  34. 34 GPT-5 mini 254
  35. 35 Claude Haiku 5.5 177
  36. 36 Claude Opus 4.7 0
#ModelScoreSettingDateSource
1 Gemini 4 Argon 1,185 Sep 2026 Official leaderboard
2 GPT-6 Astra 1,127 Sep 2026 Official leaderboard
3 GPT-6.1 Sol 973 Sep 2026 Official leaderboard
4 Claude Opus 5 938 Jul 2026 Official leaderboard
5 Claude Fable 5.1 774 Sep 2026 Official leaderboard
6 Claude Opus 5.5 723 Sep 2026 Official leaderboard
7 GPT-5.5 714 Apr 2026 Official leaderboard
8 Grok 4.7 691 Sep 2026 Official leaderboard
9 Grok 4.6 657 Aug 2026 Official leaderboard
10 GPT-5.6 Sol 638 Jul 2026 Official leaderboard
11 GPT-5.4 609 Mar 2026 Official leaderboard
12 Grok 4.5 585 Jul 2026 Official leaderboard
13 Gemini 3.1 Pro 570 Feb 2026 Official leaderboard
14 GPT-6 Sol 551 Sep 2026 Official leaderboard
15 Gemini 3.8 Flash 549 Sep 2026 Official leaderboard
16 GPT-5.2 543 Dec 2025 Official leaderboard
17 Claude Opus 4.8 522 May 2026 Official leaderboard
18 Claude Opus 4.6 496 Feb 2026 Official leaderboard
19 Gemini 3.6 Flash 480 Jul 2026 Official leaderboard
20 Gemini 3.7 Flash 471 Aug 2026 Official leaderboard
21 Claude Sonnet 5.5 460 Sep 2026 Official leaderboard
22 Gemini 3.5 Flash 456 May 2026 Official leaderboard
23 GPT-5.6 Luna 439 Jul 2026 Official leaderboard
24 GPT-5.6 Terra 433 Jul 2026 Official leaderboard
25 o3 412 Apr 2025 Official leaderboard
26 Grok 4.20 403 reasoning Feb 2026 Official leaderboard
27 GPT-5.4 Mini 392 Mar 2026 Official leaderboard
28 Claude Sonnet 4.5 344 Sep 2025 Official leaderboard
29 Claude Sonnet 5 339 Jun 2026 Official leaderboard
30 Claude Sonnet 4.6 323 Feb 2026 Official leaderboard
31 GPT-6 Luna 319 Sep 2026 Official leaderboard
32 Grok 4.3 293 Apr 2026 Official leaderboard
33 Gemini 3 Flash 277 Dec 2025 Official leaderboard
34 GPT-5 mini 254 Aug 2025 Official leaderboard
35 Claude Haiku 5.5 177 Oct 2026 Official leaderboard
36 Claude Opus 4.7 0 Apr 2026 Official leaderboard

4 models have held number one since its release in Jul 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
Gemini 4 Argon 1,185

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