MathArena - SMT 2025 — leaderboard
Metric: Accuracy (%). Source: matharena.ai. 43 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Pro (Preview) | 93.4 |
| 2 | Gemini 3 Flash | 92.92 |
| 3 | GPT-5 (High) | 91.98 |
| 4 | GPT-5.2 (High) | 91.98 |
| 5 | Step 3.5 Flash | 91.51 |
| 6 | GLM-5 | 91.04 |
| 7 | Kimi K2 (Thinking) | 91.04 |
| 8 | GPT-5.1 (High) | 91.04 |
| 9 | GLM-4.6 | 90.57 |
| 10 | Kimi K2.5 (Thinking) | 90.57 |
| 11 | GPT-5 Mini (High) | 89.15 |
| 12 | DeepSeek V3.2 Speciale | 89.15 |
| 13 | O4 Mini (High) | 88.68 |
| 14 | GPT-OSS-120B (High) | 87.74 |
| 15 | DeepSeek V3.2 (Thinking) | 87.74 |
Interactive version: theaggregate.ai/benchmark?slug=matharena-smt-2025 · How the rankings work · Data refreshed daily, snapshot 2026-07-22.