List your company
ResearchFormerly WizWand

Scientific and General Reasoning on GPQA Diamond

50.2Pass@1MulFeRL
16.9225.5634.242.84Jun 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 Learning2026.0150.2
Critique-GRPOBackbone=Qwen3-4B-Inst, Training Strategy=Reinforcement Learning-based Finetuning2026.0145.45
GRPOBackbone=Qwen3-4B-Inst, Training Strategy=Reinforcement Learning-based Finetuning2026.0143.94
Dr.GRPOBackbone=Qwen3-4B-Inst, Training Strategy=Reinforcement Learning-based Finetuning2026.0142.83
Critique-GRPOTraining Data Volume=4k, Critique Mode=CoT-Critique2025.0640.4
MulFeRLBackbone=Qwen2.5-7B-Base, Training Strategy=Multi-turn Feedback-guided Reinforcement Learning2026.0138.99
RAFTBackbone=Qwen3-4B-Inst, Training Strategy=Supervised Learning-based Finetuning2026.0137.47
Critique-GRPOBackbone=Qwen2.5-7B-Base, Training Strategy=Reinforcement Learning-based Finetuning2026.0136.97
CITL-FTBackbone=Qwen3-4B-Inst, Training Strategy=Supervised Learning-based Finetuning2026.0136.87
SFTBackbone=Qwen3-4B-Inst, Training Strategy=Supervised Learning-based Finetuning2026.0136.26
Qwen3-4B-InstBackbone=Qwen3-4B-Inst, Training Strategy=Base Model2026.0135.05
GRPOBackbone=Qwen2.5-7B-Base, Training Strategy=Reinforcement Learning-based Finetuning2026.0133.43
Dr.GRPOBackbone=Qwen2.5-7B-Base, Training Strategy=Reinforcement Learning-based Finetuning2026.0133.13
SFTBackbone=Qwen2.5-7B-Base, Training Strategy=Supervised Learning-based Finetuning2026.0128.08
Qwen2.5-7B-BaseBackbone=Qwen2.5-7B-Base, Training Strategy=Base Model2026.0127.68
CITL-FTBackbone=Qwen2.5-7B-Base, Training Strategy=Supervised Learning-based Finetuning2026.0127.07
Qwen2.5-Math-7B-BaseTraining Data Volume=None2025.0626.77
RAFTBackbone=Qwen2.5-7B-Base, Training Strategy=Supervised Learning-based Finetuning2026.0123.94
Oat-ZeroTraining Data Volume=46k2025.0623.7
SimpleRL-ZeroTraining Data Volume=46k2025.0623.2
PRIME-ZeroTraining Data Volume=46k2025.0618.2