Math
MATH-500
Problemas de matematicas de competencia (500 set).
10 modelos publicaron score
| # | Modelo | Empresa | Score |
|---|---|---|---|
| 1 | DeepSeek R1 0528 | DeepSeek | 97.3 |
| 2 | Llama 4 Behemoth | Meta | 95.0 |
| 3 | Mistral Large 3 | Mistral AI | 93.6 |
| 4 | Reka Flash 3 | Reka | 89.3 |
| 5 | Llama 4 Maverick | Meta | 87.3 |
| 6 | Nemotron 3 Nano | Nvidia | 82.9 |
| 7 | Command A | Cohere | 80.0 |
| 8 | Nova Pro | Amazon | 76.6 |
| 9 | Yi-Lightning | 01.AI | 76.4 |
| 10 | Nova Lite | Amazon | 73.3 |