Claude 3.5 Sonnet vs o1
Côte à côte. Chaque métrique. Chaque benchmark.
| Type | Claude 3.5 Sonnet | o1 |
|---|---|---|
| Provider |  Anthropic |  OpenAI |
| score moyen | 42.3 | 56.4 |
| Prix d'entrée | - | $15.00 |
| Prix de sortie | - | $60.00 |
| Fenêtre de contexte | - | 200K tokens (~100 books) |
| Sorti le | 2024-01-01 | 2024-12-17 |
| Code source ouvert | Proprietary | Proprietary |
Scores de benchmark
11 benchmarks · Claude 3.5 Sonnet: 1, o1: 9
| Benchmark | Catégorie | Claude 3.5 Sonnet | o1 |
|---|---|---|---|
| Aider — Code Editing | coding | 84.2 | 84.2 |
| Aider polyglot | coding | 51.6 | 61.7 |
| CadEval | coding | 48.0 | 56.0 |
| FrontierMath-2025-02-28-Private | math | 1.0 | 9.3 |
| GeoBench | knowledge | 62.0 | 80.0 |
| GPQA diamond | knowledge | 38.7 | 69.0 |
| Lech Mazur Writing | knowledge | 80.3 | 70.2 |
| MATH level 5 | math | 51.7 | 94.7 |
| OTIS Mock AIME 2024-2025 | math | 6.4 | 73.3 |
| SimpleBench | reasoning | 13.0 | 28.1 |
| WeirdML | coding | 31.0 | 43.8 |