List your company

Mathematical Reasoning on AIME 24 (Avg@k, Pass@k)

15.21Avg@kSFT
2.215.5858.9612.335May 8, 2026

Evaluation results

MethodLinks
SFTBackbone=Qwen2.5-7B-Instruct2026.0515.2150
RPSFTBackbone=Qwen2.5-7B-Instruct2026.0514.3843.33
IWBackbone=Qwen2.5-7B-Instruct2026.0514.1746.67
DFTBackbone=Qwen2.5-7B-Instruct2026.0512.2923.33
BaseBackbone=Qwen2.5-7B-Instruct2026.0512.0836.67
IWBackbone=Qwen2.5-3B-Instruct2026.056.6726.66
RPSFTBackbone=Qwen2.5-3B-Instruct2026.056.4633.33
SFTBackbone=Llama-3.1-8B-Instruct2026.05520
BaseBackbone=Qwen2.5-3B-Instruct2026.054.7920
RPSFTBackbone=Llama-3.1-8B-Instruct2026.054.3820
DFTBackbone=Llama-3.1-8B-Instruct2026.054.1713.33
SFTBackbone=Qwen2.5-3B-Instruct2026.053.9630
BaseBackbone=Llama-3.1-8B-Instruct2026.053.1320
IWBackbone=Llama-3.1-8B-Instruct2026.053.1316.67
DFTBackbone=Qwen2.5-3B-Instruct2026.052.7123.33