DeepSeek V3.2 Speciale vs GLM 4.5
Côte à côte. Chaque métrique. Chaque benchmark.
| Type | DeepSeek V3.2 Speciale | GLM 4.5 |
|---|---|---|
| Provider |  DeepSeek |  z-ai |
| score moyen | 78.2 | 69.2 |
| Prix d'entrée | $0.40 | $0.60 |
| Prix de sortie | $1.20 | $2.20 |
| Fenêtre de contexte | 164K tokens (~82 books) | 131K tokens (~66 books) |
| Sorti le | 2025-12-01 | 2025-07-25 |
| Code source ouvert | Open Source | Open Source |
Scores de benchmark
6 benchmarks · DeepSeek V3.2 Speciale: 6, GLM 4.5: 0
| Benchmark | Catégorie | DeepSeek V3.2 Speciale | GLM 4.5 |
|---|---|---|---|
| OpenCompass — AIME2025 | math | 96.0 | 85.8 |
| OpenCompass — GPQA-Diamond | knowledge | 86.7 | 79.5 |
| OpenCompass — HLE | knowledge | 28.6 | 16.9 |
| OpenCompass — IFEval | language | 91.7 | 85.4 |
| OpenCompass — LiveCodeBenchV6 | coding | 80.9 | 65.0 |
| OpenCompass — MMLU-Pro | knowledge | 85.5 | 82.7 |