SWE-rebench
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
SWE-rebench 9B
38.16Avg@8 BC%
9
SWE-rebench 4B
31.49Avg@8 BC%
9
SWE-rebench January 2026 (test)
52.9Resolved Rate
8
SWE-rebench 60-task Python subset v2
36.11Pass@1
7
SWE-rebench subset V2 (test)
43.7Resolved Rate
4
SWE-rebench full Python v2
22.36Pass@1
1