MathArena - HMMT Nov 2025: leaderboard
Harvard-MIT Mathematics Tournament November 2025: themed team rounds and individual subject tests for nationally competitive high school students.
Metric: Accuracy (%). Source: matharena.ai. Saturation forecast: Estimated already saturated. 23 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.2 (xHigh) | 99.17 | #110 (GPT-5.2) |
| 2 | GPT-5.2 (High) | 95.83 | #110 (GPT-5.2) |
| 3 | GLM-5 | 94.17 | #141 |
| 4 | Step 3.5 Flash | 94.17 | #270 |
| 5 | Gemini 3 Flash | 93.33 | #99 |
| 6 | Gemini 3 Pro (Preview) | 93.33 | #66 |
| 7 | Grok 4.1 Fast (Reasoning) | 93.33 | #217 (Grok 4.1 Fast) |
| 8 | DeepSeek V3.2 Speciale | 93.33 | #128 |
| 9 | GLM-4.6 | 91.67 | #252 |
| 10 | GPT-5.1 (High) | 91.67 | #135 (GPT-5.1) |
| 11 | Grok 4 Fast (Reasoning) | 90.83 | #254 (Grok 4 Fast) |
| 12 | DeepSeek V3.2 (Thinking) | 90 | #205 (DeepSeek V3.2) |
| 13 | GPT-OSS-120B (High) | 90 | #341 (GPT-OSS-120B) |
| 14 | Kimi K2 (Thinking) | 89.17 | #243 (Kimi K2) |
| 15 | Kimi K2.5 (Thinking) | 89.17 | #144 (Kimi K2.5) |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=matharena-hmmt-nov-2025 · How It Works · Data refreshed daily, snapshot 2026-10-20.