Qwen3.5-9B
2recorded runs
51.9best tok/s
1verified
Fastest GPUs for this model (best tok/s)
- RTX 306051.9
- Tenstorrent P15038.6
| GPU | Model | Quant | Backend | Ctx | tok/s | prompt tok/s | VRAM GB | Source |
|---|---|---|---|---|---|---|---|---|
| RTX 3060 · 12 GB | Qwen3.5-9B | Q4_K_M | llama.cpp b10443 | 32768 | 51.9 ✓ | 1457.9 | 5.7 | src |
| Tenstorrent P150 · 32 GB | Qwen3.5-9B | bfp_bf4 | vllm unknown | 2048 | 38.6 … | 795.6 | — | src |