Can a Mac Studio M5 Max, 128GB run local LLMs?
Yes — 33 of the 39 open models on this site fit in its 96 GB of usable memory, the strongest being Qwen3.8 27B. Whether that saves you money is a different question, and the answer is usually no.
Price$5,099
Memory128 GB, about 96 GB of it addressable by the GPU at 614 GB/s
Best model it runsQwen3.8 27B — Sonnet-class, 25 tok/s
Pay-back against the APIPays back in 64 years at 500k tokens a day
Run the numbers on this machine
What it runs
| Model | Speed | Class | Good at | Memory |
|---|---|---|---|---|
| Qwen3.8 27BQ4_K_M | 25 tok/s | Sonnet-class | 19 GB | |
| Ling 3.0 flashQ4_K_M | 26 tok/s | Haiku-class | 82 GB | |
| Qwen3.6 27BQ4_K_M | 24 tok/s | Haiku-class | 19 GB | |
| Qwen3.6 35B-A3BQ4_K_M | 73 tok/s | Haiku-class | 23 GB | |
| Muse Glimmer 30BQ4_K_M | 26 tok/s | Haiku-class | 18 GB | |
| Gemma 4 26B-A4BQ4_K_M | 54 tok/s | Haiku-class | 18 GB | |
| Qwen3.5 122B-A10BUD-Q4_K_M | 26 tok/s | Haiku-class | 79 GB | |
| Gemma 4 31B itQ4_K_M | 18 tok/s | Haiku-class | 26 GB | |
| Granite 4.2 30BQ4_K_M | 18 tok/s | Haiku-class | 26 GB | |
| GLM-4.7-FlashQ4_K_M | 52 tok/s | Haiku-class | 20 GB | |
| Gemma 4 12BQ4_K_M | 58 tok/s | Below every hosted tier | 8.0 GB | |
| Qwen3.5 9BQ4_K_M | 68 tok/s | Below every hosted tier | 6.8 GB |
21 more fit; the calculator lists them all.
The specifics
- Chip
- M5 Max — 18-core CPU / 40-core GPU
- Memory bandwidth
- 614 GB/s
- Usable by the GPU
- 96 GB — macOS lets the GPU wire roughly 75% of unified memory by default (llama.cpp discussion #2182: 66.7% at 32 GiB or less, 75% above). Raise it with `sudo sysctl iogpu.wired_limit_mb`. Memory is treated as GB throughout, which is slightly conservative.
- Power under load
- 145 W (stand in) — Apple has not published power figures for the 2026 machines yet (support articles 103253 / 102027 stop at the 2024/2025 models). Stand-in: Apple's published maximum for the previous chip (M4 Max Mac Studio, 145 W), which is a CPU+GPU stress figure, so it likely overstates inference draw.
- Availability
- Pre-order; ships 2026-09-22.
- Sources
- source 1, source 2