Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Code Generation on SWE-bench OpenHands

3.7Speedup

Hybrid Verified Decoding

0.76721.52862.293.0514May 31, 2026
Updated 1d ago

Evaluation Results

MethodLinks
2026.05
3.7
2026.05
3.31
2026.05
3.21
2026.05
2.6
2026.05
2.5
2026.05
2.11
2026.05
1.52
2026.05
1.28
2026.05
1.26
2026.05
1.09
2026.05
1.08
2026.05
0.88