Benchmark Almanac

Game Arena Chess Openings

Gemini 3 Flash 1,263 leads since Feb 2026 Not in model cards

Language models play chess against each other from varied two-move openings sampled from 20 popular human openings. The board is given as text.

Kaggle published it on 2026-02-02 with Chess in a Chess Suite. It had 11 models on 2026-11-01. Part of the Kaggle Game Arena suite. Varied openings make memorized lines less useful than in standard Chess.

Made by Kaggle and Google DeepMind Released Feb 2026 Puzzles and chess All tasks public Last model added Dec 2025 Scores refreshed 2 Nov 2026
Level
Expert Specialists with years of training, often over hours of work
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.
google-deepmind/game_arena 117 stars Leaderboard kaggle.com
11
Models scored on its own leaderboard

Gemini 3 Flash leads with 1,263 among models released by Dec 2025

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.

The newest model with a score here came out in Dec 2025, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1,263 Gemini 3 Flash
  2. 1,253 Gemini 3 Pro
  3. 934 o3
  4. 889 GPT-5.2
  5. 739 Grok 4
  6. 635 Grok 4.1 Fast
  7. 484 GPT-5 mini
  8. 252 Claude Opus 4.5
  9. 228 Claude Sonnet 4.5
  10. 102 Claude Haiku 4.5
  11. 0 DeepSeek-V3.2
  1. 1 Gemini 3 Flash 1,263
  2. 2 Gemini 3 Pro 1,253
  3. 3 o3 934
  4. 4 GPT-5.2 889
  5. 5 Grok 4 739
  6. 6 Grok 4.1 Fast reasoning 635
  7. 7 GPT-5 mini 484
  8. 8 Claude Opus 4.5 252
  9. 9 Claude Sonnet 4.5 228
  10. 10 Claude Haiku 4.5 102
  11. 11 DeepSeek-V3.2 0
#ModelScoreSettingDateSource
1 Gemini 3 Flash 1,263 Dec 2025 Official leaderboard
2 Gemini 3 Pro 1,253 Nov 2025 Official leaderboard
3 o3 934 Apr 2025 Official leaderboard
4 GPT-5.2 889 Dec 2025 Official leaderboard
5 Grok 4 739 Jul 2025 Official leaderboard
6 Grok 4.1 Fast 635 reasoning Nov 2025 Official leaderboard
7 GPT-5 mini 484 Aug 2025 Official leaderboard
8 Claude Opus 4.5 252 Nov 2025 Official leaderboard
9 Claude Sonnet 4.5 228 Sep 2025 Official leaderboard
10 Claude Haiku 4.5 102 Oct 2025 Official leaderboard
11 DeepSeek-V3.2 0 Dec 2025 Official leaderboard

Number one over time

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

Google DeepMind
Own leaderboard Not tested since Feb 2026
Gemini 3 Flash 1,263

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