community-verified · measured pool data
Can the Arc Pro B70 32GB run Qwen3.6-27B?
Yes — community-measured at 42.6 tok/s (11 runs at 4-bit).
Verified single-stream runs on the Arc Pro B70 32GB from the localmaxxing public pool. Every number on this page is measured — nothing here is estimated.
| Quantization | Median tok/s | TTFT (ms) | Peak VRAM | Max context tested | Engines |
|---|---|---|---|---|---|
| 4-bit | 42.6 | 1162.6 | 17.5 GB | 131,072 | llama.cpp, vllm |
| 5-bit | 26.6 | 311.2 | — | 204,800 | llama.cpp |