Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Code Generation on Python 10 sequential 20-step campaigns (end-of-chain)
Loading...
25.9
End pass@1
Qwen2.5-7B + Best-ckpt oracle (peak-ckpt, no gate)
4.06
9.73
15.4
21.07
Jun 17, 2026
End pass@1
Updated 1mo ago
Evaluation Results
Method
Method
Links
End pass@1
Qwen2.5-7B + Best-ckpt oracle (peak-ckpt, no gate)
Scale=7B (ref.), n=1
2026.06
25.9
Qwen2.5-7B + CARE v2 (gate + revision)
Scale=7B, n=5
2026.06
13.8
Qwen2.5-7B + Naive (end-ckpt)
Scale=7B, n=5
2026.06
11.8
Qwen2.5-3B + CARE v2 (gate + revision)
Scale=3B, n=5
2026.06
9.5
Qwen2.5-3B + Naive (end-ckpt)
Scale=3B, n=5
2026.06
4.9
Feedback
Search any
task
Search any
task