List your company

Mathematical Reasoning on MATH 500 (average@10)

48.5Average@10 ScoreQwen2.5-3B
11.5821.16530.7540.335May 26, 2026

Evaluation results

MethodLinks
Qwen2.5-3BTraining Setting=RL fine-tuned with REVERSE-MATH-augmented samples2026.0548.5
Qwen2.5-3BTraining Setting=RL fine-tuned with duplicated GSM8K samples2026.0546.2
Qwen2.5-3BTraining Setting=Untuned2026.0545.6
Llama3.1-8BTraining Setting=RL fine-tuned with duplicated GSM8K samples2026.0532.8
Llama3.2-3BTraining Setting=RL fine-tuned with REVERSE-MATH-augmented samples2026.0531.4
Llama3.1-8BTraining Setting=RL fine-tuned with REVERSE-MATH-augmented samples2026.0530.9
Llama3.2-3BTraining Setting=Untuned2026.0530.7
Llama3.2-3BTraining Setting=RL fine-tuned with duplicated GSM8K samples2026.0530.1
Llama3.1-8BTraining Setting=Untuned2026.0529.7
Mistral-7BTraining Setting=RL fine-tuned with REVERSE-MATH-augmented samples2026.0515
Mistral-7BTraining Setting=Untuned2026.0514.9
Mistral-7BTraining Setting=RL fine-tuned with duplicated GSM8K samples2026.0513