Claude Opus 4.5 vs Claude 3.5 Sonnet
Côte à côte. Chaque métrique. Chaque benchmark.
| Type | Claude Opus 4.5 | Claude 3.5 Sonnet |
|---|---|---|
| Provider |  Anthropic | Anthropic |
| score moyen | 45.4 | 42.3 |
| Prix d'entrée | $5.00 | - |
| Prix de sortie | $25.00 | - |
| Fenêtre de contexte | 200K tokens (~100 books) | - |
| Sorti le | 2025-11-24 | 2024-01-01 |
| Code source ouvert | Proprietary | Proprietary |
Scores de benchmark
11 benchmarks · Claude Opus 4.5: 11, Claude 3.5 Sonnet: 0
| Benchmark | Catégorie | Claude Opus 4.5 | Claude 3.5 Sonnet |
|---|---|---|---|
| Chatbot Arena Elo — Overall | arena | 1467.7 | 1371.4 |
| Cybench | coding | 82.0 | 17.5 |
| Fortress | safety | 13.6 | 13.0 |
| FrontierMath-2025-02-28-Private | math | 20.7 | 1.0 |
| FrontierMath-Tier-4-2025-07-01-Private | math | 4.2 | 0.1 |
| GeoBench | knowledge | 75.0 | 62.0 |
| GPQA diamond | knowledge | 81.4 | 38.7 |
| GSO-Bench | coding | 26.5 | 4.6 |
| OTIS Mock AIME 2024-2025 | math | 86.1 | 6.4 |
| SimpleBench | reasoning | 54.4 | 13.0 |
| WeirdML | coding | 63.7 | 31.0 |