Multi-Agent Pathfinding on 5-agent 8x8 evaluation set
18Valid Rate
Qwen3-4B + GRPO + LLM-as-Environment-Engineer
Updated 1mo ago
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| 18 | 10.67 | ||
2026.06 | 14 | 8.67 | |
2026.06 | 13.33 | 8 | |
2026.06 | 6.67 | 6 | |
2026.06 | 6 | 4.67 | |
2026.06 | 4.67 | 3.33 | |
2026.06 | 2 | 2 |