Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Multi-Agent Path Finding on 3-agent 8x8 (test)
Loading...
42.67
Valid Rate
Qwen3-4B + GRPO + Ours
3.8364
13.9182
24
34.0818
Jun 16, 2026
Valid Rate
Optimal Rate
Updated 1mo ago
Evaluation Results
Method
Method
Links
Valid Rate
Optimal Rate
Qwen3-4B + GRPO + Ours
Backbone=Qwen3-4B, Tra...
2026.06
42.67
24
Kimi-K2.5
2026.06
41.33
24
Grok-4.2
2026.06
36.67
21.33
Qwen3-4B + GRPO (random)
Backbone=Qwen3-4B, Tra...
2026.06
30.67
18
GPT-5.4
2026.06
26.67
15.33
Gemini-3.1-Pro
2026.06
20
10.67
Qwen3-4B (base)
Backbone=Qwen3-4B
2026.06
5.33
4.67
Feedback
Search any
task
Search any
task