Can I run this LLM?
Can a AMD Ryzen AI Max+ 395 “Strix Halo” (128GB) run Qwen3 Coder 30B-A3B (MoE)?
Qwen · Apache 2.0 · estimated for a 128.0 GB machine at 4K context.
Yes — it runs great.
Best fit: Qwen3 Coder 30B-A3B (MoE) at FP16 · 61.1 GB · 23–35 tok/s on a AMD Ryzen AI Max+ 395 “Strix Halo” (128GB).
Every quantization on this machine
| Quant | Size | Speed | Fit |
|---|---|---|---|
| Q2_K | 11.3 GB | 69–103 tok/s | Runs great |
| Q3_K_M | 14.7 GB | 61–91 tok/s | Runs great |
| Q4_K_M | 18.6 GB | 53–80 tok/s | Runs great |
| Q5_K_M | 21.7 GB | 49–73 tok/s | Runs great |
| Q6_K | 25.1 GB | 45–67 tok/s | Runs great |
| Q8_0 | 32.5 GB | 38–56 tok/s | Runs great |
| FP16 | 61.1 GB | 23–35 tok/s | Runs great |
Open the full calculator →Everything a AMD Ryzen AI Max+ 395 “Strix Halo” (128GB) runsQwen3 Coder 30B-A3B (MoE) on other machinesWhat to buy to run it →
Ran Qwen3 Coder 30B-A3B (MoE) on a AMD Ryzen AI Max+ 395 “Strix Halo” (128GB)? Add your real numbers →