MathArena - APEX Shortlist 2025 — leaderboard
Metric: Accuracy (%). Source: matharena.ai. 37 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.5 (xHigh) | 98.4 |
| 2 | Gemini 3.1 Pro (Preview) | 92.02 |
| 3 | Claude Opus 4.8 (Max) | 90.43 |
| 4 | DeepSeek V4 Flash (Max) | 89.36 |
| 5 | DeepSeek V4 Pro (Max) | 87.77 |
| 6 | Claude Opus 4.6 (High) | 86.7 |
| 7 | Gemini 3.5 Flash | 82.45 |
| 8 | GPT-5.4 (xHigh) | 81.38 |
| 9 | GPT-5.2 (High) | 79.26 |
| 10 | Kimi K2.6 (Thinking) | 77.13 |
| 11 | Step 3.7 Flash | 76.6 |
| 12 | GLM-5.1 | 71.81 |
| 13 | Step 3.5 Flash | 70.74 |
| 14 | DeepSeek V3.2 Speciale | 69.68 |
| 15 | GLM-5 | 68.62 |
Interactive version: theaggregate.ai/benchmark?slug=matharena-apex-shortlist-2025 · How the rankings work · Data refreshed daily, snapshot 2026-07-22.