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

Paper-to-Code Reproduction on PaperBench Code (dev)

78.6Final Score

Paper2Code + auto-plan & code optimized

3.51223.00642.561.994May 27, 2025Jun 27, 2025Jul 29, 2025Aug 29, 2025Sep 30, 2025Oct 31, 2025Dec 2, 2025
Updated 9d ago

Evaluation Results

MethodLinks
2025.12
78.668.215.25
2025.12
61.452.816.29
2025.05
49.6--
2025.05
48.5--
2025.05
45.1--
2025.05
44.1--
2025.05
43.4--
2025.05
17.3--
2025.05
6.4--