AIME25 leaderboard
As of November 5, 2026, among the results listed here, PrismML Ternary Bonsai 2 27B has the highest reported AIME25 score, 95, published by Prism Ml, ahead of Z.ai GLM 4.7 Flash at 91.6. 4 models have a AIME25 result in our records.
Scores as the model makers published them, not re-run by us; each row links to its announcement.
4 models ranked · 4 makers
| Rank | Model | Score | Setting | Reported by | Date |
|---|---|---|---|---|---|
| 1 | P Ternary Bonsai 2 27B | 95 | Not published | Prism Ml | September 18, 2026 (model release) |
| 2 | GLM 4.7 Flash | 91.6 | Not published | Z.ai | January 19, 2026 (model release) |
| 3 | Granite 4.2 8B | 86.7 | Not published | IBM Granite | August 31, 2026 (model release) |
| 4 | LFM2.5-2.6B | 51.9 | 2.6B | Liquid | August 11, 2026 (model release) |
How much to trust these numbers
- Source
- Reported by the model makers themselves. Not re-run by LLMBenchmark or anyone independent.
- Settings
- Settings differ between makers. Check the Setting column before comparing two scores.
- Spread
- 95 to 51.9 across 4 models.
- Freshness
- Newest result published September 18, 2026; oldest January 19, 2026.
What AIME25 measures
AIME25 asks models to solve 2025 American Invitational Mathematics Examination problems; the results do not specify what its score means. Source: math-ai/aime25 · Datasets at Hugging Face