OpenCompass Research - AIME 2025: leaderboard
OpenCompass research-table score on 2025 AIME competition math problems.
Metric: Score (%). Source: rank.opencompass.org.cn. Saturation forecast: Estimated already saturated. 48 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | DeepSeek V3.2 Speciale | 96 | #128 |
| 2 | GLM-5 | 95.8 | #141 |
| 3 | Step 3.5 Flash | 95.7 | #270 |
| 4 | GLM-4.7 | 95.4 | #192 |
| 5 | Kimi K2 (Thinking) | 94.1 | #243 (Kimi K2) |
| 6 | GPT-5 (2025-08-07) (High) | 94.1 | |
| 7 | Gemini 3 Pro (Preview) | 93.4 | #66 |
| 8 | GPT-OSS-120B (High) | 93.4 | #341 (GPT-OSS-120B) |
| 9 | DeepSeek V3.2 | 93 | #205 |
| 10 | MiMo-V2-Flash | 92.9 | #372 |
| 11 | Qwen 3.5 397B A17B | 92.3 | #146 |
| 12 | Kimi K2.5 | 91.9 | #144 |
| 13 | O4 Mini (2025-04-16) (High) | 91 | #179 (O4 Mini (2025-04-16)) |
| 14 | Qwen 3 235B A22B 2507 (Thinking) | 90.9 | #260 (Qwen 3 235B A22B 2507) |
| 15 | GPT-5 Mini (2025-08-07) (High) | 90.9 | #172 (GPT-5 Mini (2025-08-07)) |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=opencompass-research-aime-2025 · How It Works · Data refreshed daily, snapshot 2026-10-20.