List your company
ResearchFormerly WizWand

Mathematical Reasoning on HMMT25 Nov.

57.5Pass@1GPT-5.2-chat (teacher)
5.063218.676632.2945.9034May 8, 2026
Updated 5mo ago

Evaluation results

MethodLinks
GPT-5.2-chat (teacher)2026.0557.5
Qwen3-30B-A3B (teacher)Access=–2026.0555
ROPDThinking Mode=Thinking2026.0554.17
T-JudgeThinking Mode=Thinking2026.0551.25
OVDThinking Mode=Thinking2026.0550.42
GADThinking Mode=Thinking2026.0549.58
Qwen3-4B (student)Thinking Mode=Thinking2026.0548.75
ROPDThinking Mode=Non-Thinking2026.0541.67
ROPDAccess=text2026.0538.8
T-JudgeThinking Mode=Non-Thinking2026.0538.75
OVDThinking Mode=Non-Thinking2026.0537.92
ExOPDAccess=logit2026.0526.68
LOPDAccess=logit2026.0524.17
GADThinking Mode=Non-Thinking2026.0514.11
SFTAccess=text2026.058.33
Qwen3-4B (student)Thinking Mode=Non-Thinking2026.057.08
Qwen3-4B (student)Access=–2026.057.08