List your company
ResearchFormerly WizWand

Scientific and General Reasoning on MMLU Pro

68.08Pass@1MulFeRL
31.284840.837450.3959.9426Jun 3, 2025Jul 13, 2025Aug 22, 2025Oct 1, 2025Nov 10, 2025Dec 20, 2025Jan 30, 2026
Updated 8mo ago

Evaluation results

MethodLinks
MulFeRLBackbone=Qwen3-4B-Inst, Training Strategy=Multi-turn Feedback-guided Reinforcement Learning2026.0168.08
Critique-GRPOBackbone=Qwen3-4B-Inst, Training Strategy=Reinforcement Learning-based Finetuning2026.0164.67
Dr.GRPOBackbone=Qwen3-4B-Inst, Training Strategy=Reinforcement Learning-based Finetuning2026.0163.33
GRPOBackbone=Qwen3-4B-Inst, Training Strategy=Reinforcement Learning-based Finetuning2026.0162.18
SFTBackbone=Qwen3-4B-Inst, Training Strategy=Supervised Learning-based Finetuning2026.0160.46
CITL-FTBackbone=Qwen3-4B-Inst, Training Strategy=Supervised Learning-based Finetuning2026.0160.18
RAFTBackbone=Qwen3-4B-Inst, Training Strategy=Supervised Learning-based Finetuning2026.0159.53
Qwen3-4B-InstBackbone=Qwen3-4B-Inst, Training Strategy=Base Model2026.0158.56
MulFeRLBackbone=Qwen2.5-7B-Base, Training Strategy=Multi-turn Feedback-guided Reinforcement Learning2026.0154.89
Critique-GRPOBackbone=Qwen2.5-7B-Base, Training Strategy=Reinforcement Learning-based Finetuning2026.0152.34
Dr.GRPOBackbone=Qwen2.5-7B-Base, Training Strategy=Reinforcement Learning-based Finetuning2026.0151.88
GRPOBackbone=Qwen2.5-7B-Base, Training Strategy=Reinforcement Learning-based Finetuning2026.0151.1
CITL-FTBackbone=Qwen2.5-7B-Base, Training Strategy=Supervised Learning-based Finetuning2026.0147.86
SFTBackbone=Qwen2.5-7B-Base, Training Strategy=Supervised Learning-based Finetuning2026.0147.2
RAFTBackbone=Qwen2.5-7B-Base, Training Strategy=Supervised Learning-based Finetuning2026.0146.06
Qwen2.5-7B-BaseBackbone=Qwen2.5-7B-Base, Training Strategy=Base Model2026.0145.06
Critique-GRPOTraining Data Volume=4k, Critique Mode=CoT-Critique2025.0643.79
Oat-ZeroTraining Data Volume=46k2025.0641.7
Qwen2.5-Math-7B-BaseTraining Data Volume=None2025.0639.7
SimpleRL-ZeroTraining Data Volume=46k2025.0634.5
PRIME-ZeroTraining Data Volume=46k2025.0632.7