USAMO 2025 — MathArena USAMO 2025 competition problems
| # | Model | Vendor | Best score | Runs | Last seen |
|---|---|---|---|---|---|
| 1 | DeepSeek-R1 | DeepSeek | 30.1 | 4 | 2026-08-23 |
| 2 | Gemini 2.5 Pro | 24.4 | 2 | 2026-08-23 | |
| 3 | o4-mini | OpenAI | 19.1 | 2 | 2026-08-23 |
| 4 | Grok 3 | SpaceXAI | 4.8 | 2 | 2026-08-23 |
| 5 | Gemini 2.0 Flash (Reasoning) | 4.2 | 2 | 2026-08-23 | |
| 6 | Claude 3.7 Sonnet | Anthropic | 3.6 | 2 | 2026-08-23 |
| 7 | QwQ 32B Preview | Alibaba | 3 | 2 | 2026-08-23 |
| 8 | o1-pro | OpenAI | 2.8 | 2 | 2026-08-23 |
| 9 | o3-mini | OpenAI | 2.1 | 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.