LLM Stats (MathArena Apex): leaderboard
Metric: Score (%). Source: llm-stats.com. 9 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | DeepSeek V4 Pro (Max) | 90.2 |
| 2 | DeepSeek V4 Flash (Max) | 85.7 |
| 3 | Hy4 preview | 74.2 |
| 4 | DeepSeek V4 Flash (0423) | 72.1 |
| 5 | Qwen 3.7 Max | 44.5 |
| 6 | Hy3 | 38.7 |
| 7 | Seed 2.1 Turbo | 35.4 |
| 8 | Seed 2.1 Pro | 31.3 |
| 9 | Gemini 3 Pro | 23.4 |
Interactive version: theaggregate.ai/benchmark?slug=llm-stats-matharena-apex · How It Works · Data refreshed daily, snapshot 2026-09-05.