Grid Game Adaptation (10×10): leaderboard
Metric: Success rate (%; 50 random 10x10 episodes with all perturbations active: collect the key fragments and exit through the door before the step budget runs out). Source: arxiv.org. Saturation forecast: Around December 2026. 5 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Pro | 38 |
| 2 | Gemini 3 Flash | 32 |
| 3 | GPT-5 Mini | 30 |
| 4 | GPT-5.2 | 26 |
| 5 | Qwen 3 235B A22B (Non-reasoning) | 0 |
Interactive version: theaggregate.ai/benchmark?slug=grid-game-adaptation-10-10 · How It Works · Data refreshed daily, snapshot 2026-09-25.