← CanIRunLocal · All hardware · All models
Yes, if the Mac GPU limit is raised. Qwen3 235B-A22B (MoE) on M2 Ultra 192GB at Q4_K_M needs sudo sysctl iogpu.wired_limit_mb=149812 (144.3 GB needed, 144.0 GB usable), about 24 tok/s (estimate).
Hardware: M2 Ultra 192GB · Model: Qwen3 235B-A22B (MoE)
M2 Ultra 192GB has 192 GB of unified memory and 800 GB/s of memory bandwidth. The checker leaves macOS a share of unified memory, so Ollama can use 144.0 GB. Qwen3 235B-A22B (MoE) (235B, 22B active, Alibaba) at Q4_K_M is a 142 GB download and needs 144.3 GB with headroom for about 8k tokens.
| Best quant | Q4_K_M |
|---|---|
| Download size | 142 GB |
| Memory needed | 144.3 GB |
| Usable fast memory | 144.0 GB |
| Fit | fits if the Mac GPU limit is raised |
| Estimated speed | ~24 tok/s (estimate) |
| Mac GPU limit | sudo sysctl iogpu.wired_limit_mb=149812 |
| Command | ollama run qwen3:235b |
Estimates, not benchmarks. Model sizes are ollama.com download sizes; memory needed adds headroom for ~8k tokens of context; speeds are rough figures from memory bandwidth. Data updated 2026-10-23.
A project by OnDevice Labs · llms.txt