can it run
No. 8GB of memory can't hold the 21GB working set this model needs. A smaller quant or a bigger machine is required.
| model file | 16.5GB (GGUF Q4_K_M) |
| minimum memory | 21GB working set |
| this machine | 8GB LPDDR4x + 13 TOPS Hailo accelerator |
| params | 27B |
| license | apache-2.0 |
| price | $110 |
easy run: ollama pull qwen3.8:27b
runs with: ollama · llama.cpp · koboldcpp · lm studio
→ Mac mini M4 (32GB) — cheapest machine that runs it ($999)