vLLM Inference Performance on Qwen3-32B
7.641Model Load Time (s)Safetensors
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Safetensors2025.12 | 7.641 | 218.374 | 43.09 | 12,752.61 | 61.57 | |
| CryptoTensorsEncryption=Unencrypted2025.12 | 7.668 | 217.488 | 42.15 | 12,757.07 | 61.57 | |
| CryptoTensorsEncryption=Encrypted2025.12 | 111.85 | 218.584 | 42.93 | 12,795.65 | 61.57 |