AA AIME 2025: leaderboard
Artificial Analysis independent evaluation of all 30 problems from the 2025 American Invitational Mathematics Examination (olympiad-level).
Metric: Accuracy (%). Source: artificialanalysis.ai. Saturation forecast: Estimated already saturated. 270 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Best Expert | 100 | |
| 2 | GPT-5.2 (xHigh) | 99 | #110 (GPT-5.2) |
| 3 | GPT-5 Codex (High) | 98.67 | #127 (GPT-5 Codex) |
| 4 | GPT-5.2 (Medium) | 96.67 | #110 (GPT-5.2) |
| 5 | DeepSeek V3.2 Speciale | 96.67 | #128 |
| 6 | MiMo-V2-Flash (Reasoning) | 96.33 | #372 (MiMo-V2-Flash) |
| 7 | Gemini 3 Pro (Preview) (High) | 95.67 | #66 (Gemini 3 Pro (Preview)) |
| 8 | GPT-5.1 Codex (High) | 95.67 | #159 (GPT-5.1 Codex) |
| 9 | GLM-4.7 (Reasoning) | 95 | #192 (GLM-4.7) |
| 10 | Kimi K2 (Thinking) | 94.67 | #243 (Kimi K2) |
| 11 | KAT-Coder-Pro V1 | 94.67 | #324 |
| 12 | GPT-5 (High) | 94.33 | #98 (GPT-5) |
| 13 | Nova 2.0 Lite (High) | 94.33 | #417 (Nova 2.0 Lite) |
| 14 | GPT-5.1 (High) | 94 | #135 (GPT-5.1) |
| 15 | GPT-OSS-120B (High) | 93.44 | #341 (GPT-OSS-120B) |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=aa-aime-2025 · How It Works · Data refreshed daily, snapshot 2026-10-20.