Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Long-horizon agentic tasks on BrowseComp Our Settings
Loading...
62.5
Pass@1
AgentSwing
38.58
44.79
51
57.21
Mar 29, 2026
Pass@1
Updated 19d ago
Evaluation Results
Method
Method
Links
Pass@1
AgentSwing
Base Model=DeepSeek-v3.2
2026.03
62.5
AgentSwing
Base Model=Tongyi-DR-3...
2026.03
60.5
AgentSwing
Base Model=GPT-OSS-120B
2026.03
60
DeepSeek-v3.2
Context Management=Dis...
2026.03
58
Tongyi-DR-30B-A3B
Context Management=Dis...
2026.03
58
Tongyi-DR-30B-A3B
Context Management=Sum...
2026.03
55
Tongyi-DR-30B-A3B
Context Management=Kee...
2026.03
53
GPT-OSS-120B
Context Management=Kee...
2026.03
52.5
DeepSeek-v3.2
Context Management=Kee...
2026.03
52
GPT-OSS-120B
Context Management=Dis...
2026.03
50.5
DeepSeek-v3.2
Context Management=Sum...
2026.03
48.5
GPT-OSS-120B
Context Management=Sum...
2026.03
48
Tongyi-DR-30B-A3B
Context Management=Bas...
2026.03
48
DeepSeek-v3.2
Context Management=Bas...
2026.03
43.5
GPT-OSS-120B
Context Management=Bas...
2026.03
39.5
Feedback
Search any
task
Search any
task