List your company

Mathematical Reasoning on HMMT Feb 2024

19.6Average@16Ours (Shorten Prefix)
0.9845.81710.6515.483Jan 26, 2026

Evaluation results

MethodLinks
Ours (Shorten Prefix)Backbone=Qwen3-4B (non-think)2026.0119.6
GRPOBackbone=Qwen3-4B (non-think)2026.0116.7
MixedBackbone=Qwen3-4B (non-think)2026.0113.7
RTSBackbone=Qwen3-4B (non-think)2026.0113.7
Ours (Shorten Prefix)Backbone=Qwen3-4B-Base2026.0111
BaseBackbone=Qwen3-4B (non-think)2026.0110.6
RTSBackbone=Qwen3-8B-Base2026.019.5
GRPOBackbone=Qwen3-4B-Base2026.018.3
RTSBackbone=Qwen3-4B-Base2026.018
MixedBackbone=Qwen3-8B-Base2026.017.9
Ours (Shorten Prefix)Backbone=Qwen3-8B-Base2026.017.9
GRPOBackbone=Qwen3-8B-Base2026.016.4
MixedBackbone=Qwen3-4B-Base2026.016
BaseBackbone=Qwen3-4B-Base2026.012.3
BaseBackbone=Qwen3-8B-Base2026.011.7