List your company

Mathematical Reasoning on AIME 25 (avg@16)

25.3Avg@16Scaf-GRPO
-1.0125.81912.6519.481Oct 22, 2025

Evaluation results

MethodLinks
Scaf-GRPOBackbone=DeepSeek-R1-Distill-Qwen-1.5B2025.1025.3
Vanilla GRPOBackbone=DeepSeek-R1-Distill-Qwen-1.5B2025.1022.9
Scaf-GRPOBackbone=Qwen2.5-Math-7B2025.1014.6
LUFFYBackbone=Qwen2.5-Math-7B2025.1012
Oat-ZeroBackbone=Qwen2.5-Math-7B2025.1011.5
SimpleRL-ZeroBackbone=Qwen2.5-Math-7B2025.1011
Scaf-GRPOBackbone=Qwen2.5-7B2025.1011
Scaf-GRPOBackbone=Qwen2.5-Math-1.5B2025.1010.8
Vanilla GRPOBackbone=Qwen2.5-Math-7B2025.1010.8
Vanilla GRPOBackbone=Qwen2.5-7B2025.109.5
Vanilla GRPOBackbone=Qwen2.5-Math-1.5B2025.108.2
Scaf-GRPOBackbone=Llama-3.2-3B-Instruct2025.100.3
Vanilla GRPOBackbone=Llama-3.2-3B-Instruct2025.100