community-verified · measured pool data
Can the M5 Pro 24GB run Qwen3.6-35B-A3B-MTP-GGUF?
Yes — community-measured at 80.4 tok/s (2 runs at 3-bit).
Verified single-stream runs on the M5 Pro 24GB 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 |
|---|---|---|---|---|---|
| 3-bit | 80.4 | 54.1 | — | 65,536 | llama.cpp |