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

Task Performance Evaluation on PinchBench v2.0.0

89.31Best Score

All INT4

74.677278.476182.27586.0739Apr 24, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.04
89.3181.920.016958.99
2026.04
87.2585.590.015452.46
2026.04
85.7283.50.019662.22
2026.04
79.7876.950.002352.35
2026.04
78.1976.950.002357.1
2026.04
75.2473.870.002354.6