MineBench: leaderboard
Evaluates LLM spatial reasoning through Minecraft-style voxel building tasks. Head-to-head comparisons judged by community votes. Tests 3D construction understanding from text prompts.
Metric: Elo Rating. Source: minebench.ai. Source: minebench.ai. Saturation forecast: Around November 2027. 75 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 4 Argon | 2241 | #8 |
| 2 | GPT-6.1 Pro Sol | 2187 | #13 |
| 3 | Claude Opus 5.5 | 2160 | #3 |
| 4 | Claude Opus 5 | 2026 | #12 |
| 5 | Claude Sonnet 5.5 | 2015 | #11 |
| 6 | GPT-5.6 Pro Sol | 2011 | #18 |
| 7 | Claude Fable 5.1 | 1944 | #9 |
| 8 | GPT-5.5 Pro | 1937 | #17 |
| 9 | Claude Fable 5 | 1884 | #16 |
| 10 | Muse Wonder | 1876 | #15 |
| 11 | GPT-5.5 | 1861 | #31 |
| 12 | Grok 4.7 | 1853 | #25 |
| 13 | DeepSeek V4.1 Flash | 1837 | #36 |
| 14 | Grok 4.6 | 1829 | #26 |
| 15 | Gemini 3.8 Flash | 1820 | #21 |
No result here: #6 GPT-6 Astra, #7 GPT-6.1 Sol, #10 Muse Wonder Full.
Interactive version: theaggregate.ai/benchmark?slug=minebench · How It Works · Data refreshed daily, snapshot 2026-10-20.