List your company
ResearchFormerly WizWand

Mathematical Reasoning on AIME 2024 (Raw Score)

36.7Raw ScorePPO
1.96410.9822029.018Jun 11, 2025
Updated 5mo ago

Evaluation results

MethodLinks
PPOModel=Qwen2.5-Math-7B, Method type=Reinforcement Learning, Result source=-2025.0636.7
Dr. GRPO (Oat-Zero-7B)Model=Qwen2.5-Math-7B, Method type=Reinforcement Learning, Result source=-2025.0633.3
PPO-MATH7500Model=Qwen2.5-Math-7B, Method type=Reinforcement Learning, Result source=†2025.0633.3
Iterative DPOModel=Qwen2.5-Math-7B, Method type=Preference Alignment, Result source=-2025.0630
DPOModel=Qwen2.5-Math-7B, Method type=Preference Alignment, Result source=-2025.0626.7
SimPOModel=Qwen2.5-Math-7B, Method type=Preference Alignment, Result source=-2025.0626.7
Uni-DPOModel=Qwen2.5-Math-7B, Method type=Preference Alignment, Result source=-2025.0626.7
SimpleRL-Zero-7BModel=Qwen2.5-Math-7B, Method type=Reinforcement Learning, Result source=-2025.0626.7
BaselineModel=Qwen2.5-Math-7B, Method type=Base, Result source=-2025.0623.3
BaselineModel=Qwen2.5-Math-7B, Method type=Base, Result source=†2025.0623.3
Eurus2-PRIMEModel=Qwen2.5-Math-7B, Method type=Reinforcement Learning, Result source=-2025.0623.3
SFTModel=Qwen2.5-Math-7B, Method type=Supervised Fine-tuning, Result source=-2025.0620
SFTModel=Qwen2.5-Math-7B, Method type=Supervised Fine-tuning, Result source=†2025.0620
RAFTModel=Qwen2.5-Math-7B, Method type=Supervised Fine-tuning, Result source=-2025.0620
RAFTModel=Qwen2.5-Math-7B, Method type=Supervised Fine-tuning, Result source=†2025.0620
InstructModel=Llama-3.1-70B-Instruct, Result source=*2025.0616.7
OpenReasoner-Zero-7BModel=Qwen2.5-Math-7B, Method type=Reinforcement Learning, Result source=-2025.0613.3
InstructModel=Qwen2.5-Math-7B-Instruct, Result source=*2025.0613.3
GPT-4oModel=GPT-4o, Result source=*2025.069.3
Eurus2-SFTModel=Qwen2.5-Math-7B, Method type=Supervised Fine-tuning, Result source=-2025.063.3
Eurus2-SFTModel=Qwen2.5-Math-7B, Method type=Supervised Fine-tuning, Result source=*2025.063.3