Can I run this LLM?
Can a Mac Studio M3 Ultra (96GB) run Qwen3 30B-A3B (MoE)?
Qwen · Apache 2.0 · estimated for a 96.0 GB machine at 4K context.
Yes — it runs great.
Best fit: Qwen3 30B-A3B (MoE) at Q8_0 · 32.5 GB · 81–122 tok/s on a Mac Studio M3 Ultra (96GB).
Every quantization on this machine
| Quant | Size | Speed | Fit |
|---|---|---|---|
| Q2_K | 11.3 GB | 117–175 tok/s | Runs great |
| Q3_K_M | 14.7 GB | 109–164 tok/s | Runs great |
| Q4_K_M | 18.6 GB | 101–152 tok/s | Runs great |
| Q5_K_M | 21.7 GB | 96–144 tok/s | Runs great |
| Q6_K | 25.1 GB | 91–136 tok/s | Runs great |
| Q8_0 | 32.5 GB | 81–122 tok/s | Runs great |
| FP16 | 61.1 GB | 57–86 tok/s | Tight / slow |
Open the full calculator →Everything a Mac Studio M3 Ultra (96GB) runsQwen3 30B-A3B (MoE) on other machinesWhat to buy to run it →
Ran Qwen3 30B-A3B (MoE) on a Mac Studio M3 Ultra (96GB)? Add your real numbers →