International Mathematical Olympiad (IMO) 2025 — Real competition-math benchmark via MathArena; scored 0-max points scale not percentage.
| # | Model | Vendor | Best score | Runs | Last seen |
|---|---|---|---|---|---|
| 1 | GPT-5 | OpenAI | 38.1 | 2 | 2026-08-23 |
| 2 | Gemini 2.5 Pro | 31.6 | 2 | 2026-08-23 | |
| 3 | o3 | OpenAI | 16.7 | 2 | 2026-08-23 |
| 4 | o4-mini | OpenAI | 14.3 | 2 | 2026-08-23 |
| 5 | Grok 4 | SpaceXAI | 11.9 | 2 | 2026-08-23 |
| 6 | DeepSeek-R1 | DeepSeek | 6.8 | 2 | 2026-08-23 |
Best tracked score per model (default configuration; source-attributed and verification-tiered). Open a model for its full benchmark surface, provenance and pricing.