AIME 2024-2025
Math
AIME 2024-2025 is an aggregate reporting variant combining AIME problems from the 2024 and 2025 competitions. Scores are percentage accuracy under the reporting source's AVG@8 sampling protocol, averaging eight sampled solutions per problem.
Top models (higher is better)
| Model | Score |
|---|---|
| Gemini 3.1 Pro Preview | 98.1 |
| GPT-5.2 | 96.9 |
| Muse Spark | 96.9 |
| Gemini 3 Pro Preview | 96.7 |
| GPT-5.4 | 96.7 |
| Grok 4.20 | 96.5 |
| Opus 4.7 | 96.3 |
| Opus 4.6 | 95.6 |
| Gemini 3 Flash Preview | 95.6 |
| GPT-5.4 Mini | 95.6 |
| Kimi K2.5 | 95.6 |
| Opus 4.5 | 95.4 |
| Qwen3.6 Plus (2026-04-02) | 94.6 |
| GPT-5 | 93.4 |
| GLM-4.7 | 93.3 |
| GPT-5.1 | 93.3 |
| GLM 4.6 | 92.7 |
| gpt-oss-120b | 92.6 |
| Qwen3.5-Flash | 92.5 |
| Sonnet 4.6 | 92.3 |
| GLM-5.1 | 91.9 |
| Grok 4.1 Fast | 91.9 |
| GLM-5 | 91.7 |
| GPT-5 Mini | 91.5 |
| Grok 4 Fast | 91.3 |
| MiniMax M2.7 | 91.0 |
| Grok 4 | 90.6 |
| GPT-5.4 Nano | 88.8 |
| MiniMax M2.5 | 88.8 |
| Sonnet 4.5 | 88.2 |
| GLM-4.5 | 86.7 |
| o3-mini | 86.5 |
| gpt-oss-20b | 86.0 |
| Qwen3.5 Plus (2026-02-15) | 86.0 |
| Gemini 2.5 Pro Experimental 03-25 | 85.8 |
| Kimi K2 Thinking | 85.4 |
| o3 | 85.3 |
| Grok 3 Mini | 85.0 |
| DeepSeek-V3.2 | 84.6 |
| Qwen3-235B-A22B | 84.0 |