MathArena - HMMT Nov 2025 — leaderboard

Metric: Accuracy (%). Source: matharena.ai. 23 models tracked.

Top models

#ModelScore
1Human Expert100
2GPT-5.2 (xHigh)99.17
3GPT-5.2 (High)95.83
4GLM-594.17
5Step 3.5 Flash94.17
6Gemini 3 Pro (Preview)93.33
7Grok 4.1 Fast (Reasoning)93.33
8Gemini 3 Flash93.33
9DeepSeek V3.2 Speciale93.33
10GLM-4.691.67
11GPT-5.1 (High)91.67
12Grok 4 Fast (Reasoning)90.83
13GPT-OSS-120B (High)90
14DeepSeek V3.2 (Thinking)90
15GPT-5 (High)89.17

Interactive version: theaggregate.ai/benchmark?slug=matharena-hmmt-nov-2025 · How the rankings work · Data refreshed daily, snapshot 2026-07-22.