Benchmarks · reasoning
Kaggle Game Arena
Uses Kaggle's pooled Bradley-Terry leaderboard rating across strategic games. Score: Bradley-Terry rating scaled expected win rate vs Gemini 4 Argon. Collected by Jake Boggs' benchmark dashboard.
39 models · 39 results · data as of 2026-10-29 · kaggle.com ↗
Leaderboard
Gemini 4 Argon max
All results
39 rows| # | Model | Effort | Kaggle Game Arena |
|---|---|---|---|
| 1 | Gemini 4 Argon | max | 535 |
| 2 | GPT-6 Astra | max | 491 |
| 3 | Opus 5 | max | 436 |
| 4 | GPT-6.1 Sol | max | 436 |
| 5 | Gemini 3.8 Flash | max | 429 |
| 6 | Fable 5.1 | max | 418 |
| 7 | GPT-5.5 | xhigh | 407 |
| 8 | Opus 5.5 | max | 398 |
| 9 | Gemini 3.1 Pro | high | 381 |
| 10 | Gemini 3.7 Flash | max | 368 |
| 11 | GPT-5.6 Sol | max | 352 |
| 12 | GPT-5.4 | xhigh | 352 |
| 13 | Grok 4.7 | max | 335 |
| 14 | GPT-6 Sol | max | 331 |
| 15 | Gemini 3.6 Flash | xhigh | 325 |
| 16 | Sonnet 5.5 | max | 324 |
| 17 | Opus 4.8 | max | 321 |
| 18 | Gemini 3.5 Flash | xhigh | 321 |
| 19 | Grok 4.6 | max | 316 |
| 20 | GPT-5.2 | xhigh | 283 |
| 21 | Gemini 3 Flash | high | 278 |
| 22 | o3 | high | 268 |
| 23 | Opus 4.6 | max | 259 |
| 24 | Grok 4.5 | max | 258 |
| 25 | Sonnet 5 | max | 256 |
| 26 | Haiku 5.5 | max | 251 |
| 27 | Opus 4.7 | max | 209 |
| 28 | Sonnet 4.6 | max | 200 |
| 29 | GPT-5.6 Terra | max | 195 |
| 30 | Grok 4.20 | xhigh | 189 |