List your company

Mathematical Reasoning on AIME 24 (pass@1 accuracy, average pass@1 accuracy)

21.9Pass@1 AccuracySCRL
0.6846.19211.717.208Mar 20, 2026

Evaluation results

MethodLinks
SCRLBackbone=Llama-3.1-8B-Instruct2026.0321.929
ETMRBackbone=Llama-3.1-8B-Instruct2026.0316.926.2
RESTRAINBackbone=Llama-3.1-8B-Instruct2026.0316.728.4
TTRLBackbone=Llama-3.1-8B-Instruct2026.031021.2
TTRLBackbone=Llama-3.2-1B-Instruct2026.036.717.3
SCRLBackbone=Llama-3.2-1B-Instruct2026.036.723.2
Llama-3.1-8B-InstructBackbone=Llama-3.1-8B-Instruct2026.034.614
COMPASSBackbone=Llama-3.2-1B-Instruct2026.033.516.1
Llama-3.2-1B-InstructBackbone=Llama-3.2-1B-Instruct2026.031.513.1