Can the NVIDIA RTX 4080 run Llama 4 Scout 17B-16E?
No
No. Even at Q4_K_M, Llama 4 Scout 17B-16E needs about 68.3 GB, but the NVIDIA RTX 4080 has 16 GB usable. You could offload some layers to system RAM, but it will be very slow.
Try other context lengths in the calculator →Run it in the cloud
Too big for the NVIDIA RTX 4080? Run Llama 4 Scout 17B-16E on a rented GPU instead.
Rent a cloud GPU by the hour. No upfront cost, cancel any time.
Memory needed by quantization
At 8,192 tokens of context.
| Quant | Quality | Memory | Fits? | Speed |
|---|---|---|---|---|
| FP16 | Full quality | 220.2 GB | Too big | — |
| Q8_0 | Near-lossless | 118 GB | Too big | — |
| Q6_K | Excellent | 91.6 GB | Too big | — |
| Q5_K_M | Very good | 79.7 GB | Too big | — |
| Q4_K_M | Good, most popular | 68.3 GB | Too big | — |
| Q3_K_M | Noticeably degraded | 55.5 GB | Too big | — |
| Q2_K | Heavily degraded | 43.1 GB | Too big | — |
How to run Llama 4 Scout 17B-16E anyway
On a rented cloud GPU with enough memory, the same command you'd use locally works:
ollama run llama4:scoutAt Q4_K_M it needs about 68.3 GB, so pick a GPU with at least 72 GB.