List your company

Mathematical Reasoning on MATH-500 (Acc, Tok, ReL Metrics)

90AccuracyOurs
74.81678.75882.786.642Feb 1, 2026

Evaluation results

MethodLinks
OursBackbone=Qwen2.5-7B2026.02904,196140.6
Ours (Base)Backbone=Qwen2.5-7B2026.0287.44,253161.6
INFTYTHINKBackbone=Qwen2.5-7B2026.0286.25,050184.8
LightThinkerBackbone=Qwen2.5-7B2026.0284.84,321145.2
SepLLMBackbone=Qwen2.5-7B2026.0284.24,244144.3
H2OBackbone=Qwen2.5-7B2026.0282.64,389147.1
OursBackbone=Qwen2.5-1.5B2026.0281.23,812125.8
Ours (Base)Backbone=Qwen2.5-1.5B2026.0278.83,958139.4
INFTYTHINKBackbone=Qwen2.5-1.5B2026.02784,750170.2
LightThinkerBackbone=Qwen2.5-1.5B2026.0277.64,060134.8
SepLLMBackbone=Qwen2.5-1.5B2026.02763,890130.3
H2OBackbone=Qwen2.5-1.5B2026.0275.43,835128.5