Can the NVIDIA RTX 5090 run Mistral Small 24B?
Yes
QuantQ8_0Near-lossless
Memory27 GBof 32 GB
Speed~43tokens/s
Yes. The NVIDIA RTX 5090 can run Mistral Small 24B fully in VRAM at Q8_0 (near-lossless), using about 27 GB of the 32 GB available, at roughly 43 tokens/second.
Try other context lengths in the calculator →Memory needed by quantization
At 8,192 tokens of context.
| Quant | Quality | Memory | Fits? | Speed |
|---|---|---|---|---|
| FP16 | Full quality | 49.1 GB | Too big | — |
| Q8_0 | Near-lossless | 27 GB | Fits | ~43 tok/s |
| Q6_K | Excellent | 21.3 GB | Fits | ~56 tok/s |
| Q5_K_M | Very good | 18.7 GB | Fits | ~64 tok/s |
| Q4_K_M | Good, most popular | 16.2 GB | Fits | ~75 tok/s |
| Q3_K_M | Noticeably degraded | 13.5 GB | Fits | ~93 tok/s |
| Q2_K | Heavily degraded | 10.8 GB | Fits | ~121 tok/s |
How to run Mistral Small 24B on the NVIDIA RTX 5090
With Ollama installed, one command downloads and starts it:
ollama run mistral-small:24bOllama's default download is Q4_K_M (about 16.2 GB here). You have room for Q8_0, which is near-lossless: pick that GGUF in LM Studio or from Hugging Face.
GPUs that run Mistral Small 24B comfortably
- NVIDIA RTX 3090 (24 GB)
- AMD Radeon RX 7900 XTX (24 GB)
- NVIDIA RTX 4090 (24 GB)
- Apple M2 Pro (32GB)
- Apple M4 Pro (48GB)
Models that run well on the NVIDIA RTX 5090
Run it in the cloud
Don't want to buy hardware? Run Mistral Small 24B on a rented GPU today.
Rent a cloud GPU by the hour. No upfront cost, cancel any time.