R1 vs o3
Côte à côte. Chaque métrique. Chaque benchmark.
| Type | R1 | o3 |
|---|---|---|
| Provider |  DeepSeek |  OpenAI |
| score moyen | 45.1 | 55.2 |
| Prix d'entrée | $0.70 | $2.00 |
| Prix de sortie | $2.50 | $8.00 |
| Fenêtre de contexte | 64K tokens (~32 books) | 200K tokens (~100 books) |
| Sorti le | 2025-01-20 | 2025-04-16 |
| Code source ouvert | Open Source | Proprietary |
Scores de benchmark
12 benchmarks · R1: 0, o3: 12
| Benchmark | Catégorie | R1 | o3 |
|---|---|---|---|
| Aider polyglot | coding | 56.9 | 81.3 |
| ARC-AGI | reasoning | 15.8 | 60.8 |
| ARC-AGI-2 | reasoning | 1.3 | 6.5 |
| DeepResearch Bench | knowledge | 35.1 | 46.6 |
| Fiction.LiveBench | knowledge | 69.4 | 88.9 |
| GPQA diamond | knowledge | 62.3 | 75.8 |
| Lech Mazur Writing | knowledge | 83.0 | 83.9 |
| MATH level 5 | math | 93.0 | 97.8 |
| OTIS Mock AIME 2024-2025 | math | 53.3 | 83.9 |
| SimpleBench | reasoning | 17.1 | 43.7 |
| SimpleQA Verified | knowledge | 27.4 | 53.0 |
| WeirdML | coding | 36.5 | 52.4 |