Claude Haiku 4.5 vs GPT-5 Mini
Côte à côte. Chaque métrique. Chaque benchmark.
| Type | Claude Haiku 4.5 | GPT-5 Mini |
|---|---|---|
| Provider |  Anthropic |  OpenAI |
| score moyen | 37.1 | 56.0 |
| Prix d'entrée | $1.00 | $0.25 |
| Prix de sortie | $5.00 | $2.00 |
| Fenêtre de contexte | 200K tokens (~100 books) | 400K tokens (~200 books) |
| Sorti le | 2025-10-15 | 2025-08-07 |
| Code source ouvert | Proprietary | Proprietary |
Scores de benchmark
10 benchmarks · Claude Haiku 4.5: 1, GPT-5 Mini: 9
| Benchmark | Catégorie | Claude Haiku 4.5 | GPT-5 Mini |
|---|---|---|---|
| ARC-AGI | reasoning | 47.7 | 54.3 |
| ARC-AGI-2 | reasoning | 4.0 | 4.4 |
| FrontierMath-2025-02-28-Private | math | 5.9 | 27.2 |
| FrontierMath-Tier-4-2025-07-01-Private | math | 2.1 | 6.3 |
| GPQA diamond | knowledge | 61.6 | 66.7 |
| MATH level 5 | math | 96.4 | 97.8 |
| OTIS Mock AIME 2024-2025 | math | 66.6 | 86.7 |
| SimpleQA Verified | knowledge | 5.9 | 21.0 |
| Terminal Bench | coding | 35.5 | 34.8 |
| WeirdML | coding | 45.4 | 52.7 |