Can the NVIDIA RTX 3080 10GB run Mixtral 8x7B?
No
No. Even at Q4_K_M, Mixtral 8x7B needs about 30 GB, but the NVIDIA RTX 3080 10GB has 10 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 3080 10GB? Run Mixtral 8x7B 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 | 95.1 GB | Too big | — |
| Q8_0 | Near-lossless | 51.3 GB | Too big | — |
| Q6_K | Excellent | 40 GB | Too big | — |
| Q5_K_M | Very good | 34.9 GB | Too big | — |
| Q4_K_M | Good, most popular | 30 GB | Too big | — |
| Q3_K_M | Noticeably degraded | 24.5 GB | Too big | — |
| Q2_K | Heavily degraded | 19.2 GB | Too big | — |
How to run Mixtral 8x7B anyway
On a rented cloud GPU with enough memory, the same command you'd use locally works:
ollama run mixtral:8x7bAt Q4_K_M it needs about 30 GB, so pick a GPU with at least 32 GB.
GPUs that run Mixtral 8x7B comfortably
- NVIDIA RTX 5090 (32 GB)
- Apple M4 Pro (48GB)
- Apple M3 Max (64GB)
- Apple M4 Max (128GB)
- Apple M2 Ultra (192GB)