MathArena - BRUMO 2025: leaderboard
Brown University Math Olympiad 2025: algebra, combinatorics, geometry, and number theory across individual, team, and guts rounds.
Metric: Accuracy (%). Source: matharena.ai. Saturation forecast: Estimated already saturated. 45 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Flash | 100 | #99 |
| 2 | Step 3.5 Flash | 100 | #270 |
| 3 | GLM-5 | 99.17 | #141 |
| 4 | DeepSeek V3.2 Speciale | 99.17 | #128 |
| 5 | Gemini 3 Pro (Preview) | 98.33 | #66 |
| 6 | Kimi K2.5 (Thinking) | 98.33 | #144 (Kimi K2.5) |
| 7 | GPT-5.2 (High) | 98.33 | #110 (GPT-5.2) |
| 8 | Grok 4.1 Fast (Reasoning) | 97.5 | #217 (Grok 4.1 Fast) |
| 9 | DeepSeek V3.2 (Thinking) | 96.67 | #205 (DeepSeek V3.2) |
| 10 | Grok 4 Fast (Reasoning) | 95.83 | #254 (Grok 4 Fast) |
| 11 | O3 (High) | 95.83 | #124 (O3) |
| 12 | DeepSeek V3.2 Exp (Thinking) | 95.83 | #234 (DeepSeek V3.2 Exp) |
| 13 | Grok 4 | 95 | #181 |
| 14 | GLM-4.6 | 94.17 | #252 |
| 15 | Kimi K2 (Thinking) | 93.33 | #243 (Kimi K2) |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=matharena-brumo-2025 · How It Works · Data refreshed daily, snapshot 2026-10-20.