List your company

Mathematical Reasoning on MATH500 (Accuracy and Resource Usage)

12Memory Usage RatioESC
8.4832.245679.76Jan 29, 2026

Evaluation results

MethodLinks
ESCBackbone model=Qwen3-4B2026.01128892.11
ESCBackbone model=Deepseek-R1-Distill-Qwen-7B2026.01158494.02
DSCBackbone model=Qwen3-4B2026.011613591.17
T2Backbone model=Qwen2.5-7B2026.01184878.68
ACBackbone model=Llama3.1-8B2026.011817157.9
ACBackbone model=Qwen3-4B2026.011815692.17
T2Backbone model=Deepseek-R1-Distill-Qwen-7B2026.01193794.32
T2Backbone model=Qwen3-4B2026.01204591.96
T2Backbone model=Llama3.1-8B2026.01234158.46
ACBackbone model=Qwen2.5-7B2026.012414978.26
DSCBackbone model=Llama3.1-8B2026.012817557.38
ACBackbone model=Deepseek-R1-Distill-Qwen-7B2026.012818993.7
DeepConfBackbone model=Qwen3-4B2026.012927791.87
ESCBackbone model=Llama3.1-8B2026.013414058.65
DSCBackbone model=Qwen2.5-7B2026.014414378.33
DeepConfBackbone model=Llama3.1-8B2026.014423758.12
DeepConfBackbone model=Qwen2.5-7B2026.014534578.82
DSCBackbone model=Deepseek-R1-Distill-Qwen-7B2026.014521391.04
ESCBackbone model=Qwen2.5-7B2026.015227778.58
DeepConfBackbone model=Deepseek-R1-Distill-Qwen-7B2026.015338693.45
Std-PT.Backbone model=Qwen2.5-7B2026.0110010079.37
Std-PT.Backbone model=Llama3.1-8B2026.0110010058.89
Std-PT.Backbone model=Deepseek-R1-Distill-Qwen-7B2026.0110010094.24
Std-PT.Backbone model=Qwen3-4B2026.0110010092.52