List your company

Software Engineering on PaperBench

66.8ScoreClaude Sonnet 4.5
8.14423.37238.653.828Mar 23, 2026

Evaluation results

MethodLinks
Claude Sonnet 4.5SDK=v1.11.0, Agent Configuration=SINGLE-AGENT + CAID2026.0366.83,883.912.6
Claude Sonnet 4.5SDK=v1.11.0, Agent Configuration=CAID, Number of Engineers=22026.0363.32,080.49.3
Claude Sonnet 4.5SDK=v1.11.0, Agent Configuration=Single-Agent2026.0357.21,803.53.3
GLM 4.7SDK=v1.11.0, Agent Configuration=SINGLE-AGENT + CAID2026.0348.52,626.97.5
GLM 4.7SDK=v1.11.0, Agent Configuration=CAID, Number of Engineers=22026.0345.41,449.44.7
GLM 4.7SDK=v1.11.0, Agent Configuration=Single-Agent2026.03381,177.62.8
MiniMax 2.5SDK=v1.11.0, Agent Configuration=CAID, Number of Engineers=22026.0336.72,999.42.6
MiniMax 2.5SDK=v1.11.0, Agent Configuration=SINGLE-AGENT + CAID2026.0336.75,524.73.7
MiniMax 2.5SDK=v1.11.0, Agent Configuration=Single-Agent2026.0310.42,525.31.1