Can I run this LLM?
Can a AMD Ryzen AI Max+ 395 “Strix Halo” (64GB) run Gemma 4 31B?
Gemma · Apache 2.0 · estimated for a 64.0 GB machine at 4K context.
Yes — it runs great.
Best fit: Gemma 4 31B at Q2_K · 11.8 GB · 12–19 tok/s on a AMD Ryzen AI Max+ 395 “Strix Halo” (64GB).
Every quantization on this machine
| Quant | Size | Speed | Fit |
|---|---|---|---|
| Q2_K | 11.8 GB | 12–19 tok/s | Runs great |
| Q3_K_M | 14.7 GB | 10–15 tok/s | Tight / slow |
| Q4_K_M | 18.3 GB | 8–13 tok/s | Tight / slow |
| Q5_K_M | 21.7 GB | 7–11 tok/s | Tight / slow |
| Q6_K | 25.2 GB | 6–9 tok/s | Tight / slow |
| Q8_0 | 32.6 GB | 5–7 tok/s | Tight / slow |
| FP16 | 61.4 GB | 3–4 tok/s | Won't fit |
Open the full calculator →Everything a AMD Ryzen AI Max+ 395 “Strix Halo” (64GB) runsGemma 4 31B on other machinesWhat to buy to run it →
Ran Gemma 4 31B on a AMD Ryzen AI Max+ 395 “Strix Halo” (64GB)? Add your real numbers →