R1 vs o1
Lado a lado. Cada métrica. Cada benchmark.
| Tipo | R1 | o1 |
|---|---|---|
| Provider |  DeepSeek |  OpenAI |
| puntuación promedio | 45.1 | 56.4 |
| Precio de entrada | $0.70 | $15.00 |
| Precio de salida | $2.50 | $60.00 |
| Ventana de contexto | 64K tokens (~32 books) | 200K tokens (~100 books) |
| Publicado el | 2025-01-20 | 2024-12-17 |
| Código abierto | Open Source | Proprietary |
Puntuaciones de benchmark
9 benchmarks · R1: 1, o1: 8
| Benchmark | Categoría | R1 | o1 |
|---|---|---|---|
| Aider polyglot | coding | 56.9 | 61.7 |
| ARC-AGI | reasoning | 15.8 | 30.7 |
| Fiction.LiveBench | knowledge | 69.4 | 83.3 |
| GPQA diamond | knowledge | 62.3 | 69.0 |
| Lech Mazur Writing | knowledge | 83.0 | 70.2 |
| MATH level 5 | math | 93.0 | 94.7 |
| OTIS Mock AIME 2024-2025 | math | 53.3 | 73.3 |
| SimpleBench | reasoning | 17.1 | 28.1 |
| WeirdML | coding | 36.5 | 43.8 |