DeepSeek V3.2 vs gpt-oss-20b (free)
Côte à côte. Chaque métrique. Chaque benchmark.
| Type | DeepSeek V3.2 | gpt-oss-20b (free) |
|---|---|---|
| Provider |  DeepSeek |  OpenAI |
| score moyen | 53.0 | 66.4 |
| Prix d'entrée | $0.26 | $0.00 |
| Prix de sortie | $0.38 | $0.00 |
| Fenêtre de contexte | 164K tokens (~82 books) | 131K tokens (~66 books) |
| Sorti le | 2025-12-01 | 2025-08-05 |
| Code source ouvert | Open Source | Open Source |
Scores de benchmark
9 benchmarks · DeepSeek V3.2: 9, gpt-oss-20b (free): 0
| Benchmark | Catégorie | DeepSeek V3.2 | gpt-oss-20b (free) |
|---|---|---|---|
| Artificial Analysis — Agentic Index | speed | 52.9 | 27.6 |
| Artificial Analysis — Coding Index | speed | 36.7 | 18.5 |
| Artificial Analysis — Quality Index | speed | 41.7 | 24.5 |
| OpenCompass — AIME2025 | math | 93.0 | 87.9 |
| OpenCompass — GPQA-Diamond | knowledge | 84.6 | 68.9 |
| OpenCompass — HLE | knowledge | 23.2 | 11.6 |
| OpenCompass — IFEval | language | 89.7 | 88.9 |
| OpenCompass — LiveCodeBenchV6 | coding | 75.4 | 68.4 |
| OpenCompass — MMLU-Pro | knowledge | 85.8 | 72.8 |