LMGame-Bench Tetris — leaderboard

LLM game-playing benchmark: Tetris testing spatial reasoning and piece placement optimization.

Metric: Score. Source: huggingface.co. Status: saturation imminent. 25 models tracked.

Top models

#ModelScore
1Grok 4125.7
2GPT-5 (High)84.3
3O3 (2025-04-16)42
4O1 (2024-12-17)35
5DeepSeek R1 052833.7
6O4 Mini (2025-04-16)25.3
7Gemini 2.5 Pro (Preview 05-06)23.3
8Grok 3 Mini Beta21.3
9Claude Opus 4 (20250514)20
10GLM-4.519.7
11Claude Sonnet 4 (20250514)19.3
12Kimi K2 (0711)17
13Claude 3.7 Sonnet (20250219)16.3
14Gemini 2.5 Flash (Preview 04-17)16.3
15Claude 3.5 Sonnet (20241022)14.7

Interactive version: theaggregate.ai/benchmark?slug=lmgame-bench-tetris · How the rankings work · Data refreshed daily, snapshot 2026-07-22.