List your company

Reasoning Performance Aggregation on Aggregate Benchmarks Code Math

56.2Code Component AverageSAGE
46.52849.03951.5554.061Mar 16, 2026

Evaluation results

MethodLinks
SAGEBackbone=Qwen-3-4B-Base2026.0356.255.855.9
SAGEBackbone=Qwen-2.5-7B-Instruct2026.0355.547.550.1
AZRBackbone=Qwen-2.5-7B-Instruct2026.0355.246.849.6
AZRBackbone=Qwen-3-4B-Base2026.0355.246.749.5
MAEBackbone=Qwen-3-4B-Base2026.0355.253.954.4
MAEBackbone=Qwen-2.5-7B-Instruct2026.0354.94548.3
Base ModelBackbone=Qwen-3-4B-Base2026.0354.556.355.7
Base ModelBackbone=Qwen-2.5-7B-Instruct2026.035245.447.6
SAGEBackbone=Qwen-2.5-3B-Instruct2026.0349.438.342
AZRBackbone=Qwen-2.5-3B-Instruct2026.0348.435.739.9
MAEBackbone=Qwen-2.5-3B-Instruct2026.0348.436.640.5
Base ModelBackbone=Qwen-2.5-3B-Instruct2026.0346.937.240.4