MineExplorer: leaderboard
Multimodal large language models (MLLMs) have shown strong capabilities in perception, reasoning, and action generation. However, their ability to sustain exploration in dynamic open worlds remains unclear. Existing embodied and game-based benchmarks often compress interaction into short-horizon tasks or entangle success with domain-specific game mechanics. In this paper, we introduce MineExplorer
Metric: Overall TSR (self-reported). Source: benchmarklist.com. Saturation forecast: Around June 2027. 18 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.6 | 41.08 | #63 |
| 2 | Gemini 3.1 Pro (Preview) | 37.02 | #58 |
| 3 | Claude Opus 4.5 | 27.31 | #83 |
| 4 | GPT-5.2 | 21.28 | #110 |
| 5 | GLM-5V Turbo | 19.93 | #166 |
| 6 | Claude Sonnet 4.5 | 18.82 | #142 |
| 7 | GPT-5.4 | 18.08 | #79 |
| 8 | Claude Haiku 4.5 | 15.5 | #280 |
| 9 | Seed 2.0 Pro | 15.5 | #118 |
| 10 | Gemini 2.5 Flash | 15.38 | #245 |
| 11 | Gemini 2.5 Pro | 14.51 | #152 |
| 12 | GPT-4.1 | 12.18 | #248 |
| 13 | Qwen 3 VL 235B A22B Instruct | 10.58 | #273 |
| 14 | Qwen 3 VL 32B Instruct | 10.09 | #283 |
| 15 | Kimi K2.6 | 9.96 | #104 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=mineexplorer · How It Works · Data refreshed daily, snapshot 2026-10-20.