Multi-Agent Path Finding on MAPF-FrozenLake 4-agent (9x9)
25.33Valid RateQwen3-4B + GRPO + LLM-as-Environment-Engineer
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Qwen3-4B + GRPO + LLM-as-Environment-EngineerBackbone=Qwen3-4B, Training=GRPO, Environment Design=Ours2026.06 | 25.33 | 16.67 | |
| Kimi-K2.52026.06 | 20.67 | 14 | |
| Qwen3-4B + GRPO (random)Backbone=Qwen3-4B, Training=GRPO, Config=single randomly sampled Round-02026.06 | 19.33 | 10.67 | |
| GPT-5.42026.06 | 9.33 | 5.33 | |
| Grok-4.22026.06 | 8.67 | 6.67 | |
| Gemini-3.1-Pro2026.06 | 4.67 | 2 | |
| Qwen3-4B (base)Backbone=Qwen3-4B, Status=untrained2026.06 | 0 | 0 |