Apple M4 Max (40-core GPU, 128GB) × Llama 4 Scout (17B active, 16 experts)
Can the Apple M4 Max (40-core GPU, 128GB) run Llama 4 Scout (17B active, 16 experts)?
Yes, usable
Yes. The Apple M4 Max (40-core GPU, 128GB) can run Llama 4 Scout (17B active, 16 experts) at up to Q6_K (needs ~95.0 GB of 96.0 GB usable), at an estimated ~18.6 tokens/sec at Q4_K_M.
Speed is estimated from memory bandwidth calibrated with a measured 83.06 tok/s Llama 2 7B Q4_0 result (source). No direct public benchmark of this exact pair is recorded yet.
By quantization (8K context)
| Quant | Memory needed | Fits? | Est. tok/s |
|---|---|---|---|
| Q3_K_M | 57.1 GB | ✅ | ~22.9 |
| Q4_K_M | 70.5 GB | ✅ | ~18.6 |
| Q5_K_M | 82.4 GB | ✅ | ~15.9 |
| Q6_K | 95.0 GB | ✅ | ~13.8 |
| Q8_0 | 122 GB | ❌ | — |
| F16 | 227 GB | ❌ | — |
Fastest hardware for Llama 4 Scout (17B active, 16 experts)
- Apple M5 Max (40-core GPU, 128GB) — ~26.9 tok/s at Q4_K_M
- Apple M2 Ultra (76-core GPU, 192GB) — ~21.1 tok/s at Q4_K_M
- Apple M3 Ultra (80-core GPU, 512GB) — ~20.7 tok/s at Q4_K_M
- NVIDIA DGX Spark — ~12.8 tok/s at Q4_K_M
No suitable GPU? Rent one by the hour on Vast.ai or RunPod (affiliate links).
Get new models & benchmarks in your inbox
No spam. Unsubscribe anytime by replying to any email.