Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Long-horizon Agentic Reasoning on BrowseComp (Pass@3)
Loading...
34
Pass@3
AgentCPM-Explore + Long-context RL
26.72
28.61
30.5
32.39
Jun 17, 2026
Pass@3
Updated 1mo ago
Evaluation Results
Method
Method
Links
Pass@3
AgentCPM-Explore + Long-context RL
Training Steps=50 step...
2026.06
34
AgentCPM-Explore + Long-context RL
Training Steps=25 step...
2026.06
29
AgentCPM-Explore
Data Recipe=Base Agent...
2026.06
27
Feedback
Search any
task
Search any
task
