List your company

Mathematical Reasoning on OlympiadBench (Avg@16, Pass@16)

36.36Avg@16PriFT-mass
16.91221.96127.0132.059Jun 8, 2026

Evaluation results

MethodLinks
PriFT-massBackbone=Qwen3-Base (8B)2026.0636.3664
PriFT-probBackbone=Qwen2.5-Math (7B)2026.0636.1465.78
PriFT-massBackbone=Qwen2.5-Math (7B)2026.0635.7561.78
TALRBackbone=Qwen2.5-Math (7B)2026.0634.7159.85
DFTBackbone=Qwen2.5-Math (7B)2026.0634.556
PriFT-probBackbone=Qwen3-Base (8B)2026.0634.263.85
ASFTBackbone=Qwen2.5-Math (7B)2026.0633.9563.11
DFTBackbone=Qwen3-Base (8B)2026.0633.3458.07
IDFTBackbone=Qwen3-Base (8B)2026.0633.249.33
ASFTBackbone=Qwen3-Base (8B)2026.0632.4163.41
IDFTBackbone=Qwen2.5-Math (7B)2026.0631.7149.78
TALRBackbone=Qwen3-Base (8B)2026.0631.758.67
OriginalBackbone=Qwen3-Base (8B)2026.0627.2366.07
SFTBackbone=Qwen3-Base (8B)2026.061954.07
EAFTBackbone=Qwen3-Base (8B)2026.0618.953.93
EAFTBackbone=Qwen2.5-Math (7B)2026.0618.8554.67
SFTBackbone=Qwen2.5-Math (7B)2026.0618.7954.52
OriginalBackbone=Qwen2.5-Math (7B)2026.0617.6658.67