List your company

Mathematical Reasoning on DART-Math (DM-1 to DM-5)

84.8DM-1 ScoreFlexiDepth
-3.39219.50442.465.296Jun 4, 2026

Evaluation results

MethodLinks
FlexiDepthBase Model=Qwen3-8B, p@=52026.0684.859.219.46.86.6
FlexiDepthBase Model=Qwen3-8B, p@=42026.0681.255.618.45.24.8
FlexiDepthBase Model=Qwen3-8B, p@=32026.0675.850.616.44.22.8
FlexiDepthBase Model=Qwen3-8B, p@=22026.066742.213.83.22
POLARBase Model=Qwen3-8B, p@=52026.0655.644.825.418.817.8
POLARBase Model=Qwen3-8B, p@=42026.0653.44424.818.417.6
POLARBase Model=Qwen3-8B, p@=32026.0650.84122.21817.2
FlexiDepthBase Model=Qwen3-8B, p@=12026.0649.629.48.81.81
Base (sampling)Base Model=Qwen3-8B, p@=52026.0648.427.813.613.410.6
POLARBase Model=Qwen3-8B, p@=22026.0648.433.221.21614.6
Base (sampling)Base Model=Qwen3-8B, p@=42026.064726.612.612.810.2
Base (sampling)Base Model=Qwen3-8B, p@=32026.0645.82612.411.49.8
POLARBase Model=Qwen3-8B, p@=12026.0643.230.420.813.514.9
Base (sampling)Base Model=Qwen3-8B, p@=22026.064225.611.610.68.8
BaseBase Model=Qwen3-8B, Temperature (τ)=0, p@=12026.0641.628.8191213
Base (sampling)Base Model=Qwen3-8B, p@=12026.0637239.29.67
DR.LLMBase Model=Qwen3-8B, p@=22026.062.40.60.80.22.4
DR.LLMBase Model=Qwen3-8B, p@=32026.062.40.60.80.22.4
DR.LLMBase Model=Qwen3-8B, p@=42026.062.40.60.80.22.8
DR.LLMBase Model=Qwen3-8B, p@=52026.062.410.80.22.8
DR.LLMBase Model=Qwen3-8B, p@=12026.0620.60.60.22
ShortGPTBase Model=Qwen3-8B, p@=12026.0600000
ShortGPTBase Model=Qwen3-8B, p@=22026.0600000
ShortGPTBase Model=Qwen3-8B, p@=32026.0600000
ShortGPTBase Model=Qwen3-8B, p@=42026.0600000
ShortGPTBase Model=Qwen3-8B, p@=52026.0600000
MindSkipBase Model=Qwen3-8B, p@=12026.0600000
MindSkipBase Model=Qwen3-8B, p@=22026.0600000
MindSkipBase Model=Qwen3-8B, p@=32026.0600000
MindSkipBase Model=Qwen3-8B, p@=42026.0600000
MindSkipBase Model=Qwen3-8B, p@=52026.0600000