MathArena - HMMT Feb 2026 — leaderboard

Metric: Accuracy (%). Source: matharena.ai. 30 models tracked.

Top models

#ModelScore
1Human Expert100
2GPT-5.5 (xHigh)98.48
3GPT-5.4 (xHigh)97.73
4GPT-5.2 (High)96.97
5Claude Opus 4.6 (High)96.21
6Gemini 3.5 Flash95.45
7Claude Opus 4.8 (Max)95.45
8Gemini 3.1 Pro (Preview)94.7
9Kimi K2.6 (Thinking)94.7
10DeepSeek V4 Pro (Max)93.94
11DeepSeek V4 Flash (Max)93.94
12Claude Opus 4.7 (xHigh)93.94
13GLM-5.292.42
14GLM-5.189.39
15Gemini 3 Flash89.39

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