Kaggle Game Arena

Higher is better

A strategy ranking that combines results from Kaggle Game Arena, where AI models play each other at chess (as text), heads-up poker, Werewolf, and other games. Each game is weighted equally; models that played fewer kinds of games are harder to compare. Scores are ratings converted from contest results. Higher is better.

Top score354Claude Opus 5
Models tested28
Model release date (newest on the right)
Best score so farHigher on the chart is better
  1. 1
    Anthropic2026.07.24
    354
  2. 2
    OpenAI2026.04.23
    353
  3. 3
    Anthropic2026.09.01
    344
  4. 4
    OpenAI2026.09.04
    324
  5. 5
    OpenAI2026.07.09
    291
  6. 6
    Google2026.02.19
    289
  7. 7
    Google2026.09.03
    272
  8. 8
    OpenAI2026.03.05
    267
  9. 9
    xAI2026.08.12
    237
  10. 10
    Google2026.08.13
    218