Can I Run Codestral on RTX 5090 Desktop (32 GB VRAM, 64 GB RAM)?

Superseded model. Codestral has been superseded by Devstral. This page is kept for reference; the newer family is a better starting point. View Devstral →

Written by Jakub Rusinowski · Last updated May 29, 2024

Yes

Yes — Codestral 22B at Q8_0 needs about 26.3 GB of the 32 GB usable on RTX 5090 Desktop (32 GB VRAM, 64 GB RAM) (~5.7 GB spare), at ~48.8 tok/s (estimated), with room for about 32,768 tokens of context.

Confidence: high · Recommended quantization: Q8_0 · Estimated speed: ~48.8 tok/s

RTX 5090 Desktop (32 GB VRAM, 64 GB RAM) — what it gives a model

Usable memory for models	32 GB
Memory bandwidth	1792 GB/s
Form factor	Desktop
Operating system	Windows or Linux
Memory upgradeable	Yes

Codestral on RTX 5090 Desktop (32 GB VRAM, 64 GB RAM): memory by quantization

Quant	Memory needed	Fits 32 GB?	Max context	Est. speed	Download
F16	47.1 GB	✗ No	—	—	44.4 GB
Q8_0	26.3 GB	✓ Yes	32K	~48.8 tok/s	23.6 GB
Q6_K	20.9 GB	✓ Yes	32K	~60.6 tok/s	18.2 GB
Q5_K_M	18.4 GB	✓ Yes	32K	~68.2 tok/s	15.7 GB
Q4_K_M	16.1 GB	✓ Yes	32K	~77.4 tok/s	13.4 GB
Q3_K_M	12.1 GB	✓ Yes	32K	~100.1 tok/s	9.5 GB
Q2_K	10 GB	✓ Yes	32K	~119.4 tok/s	7.3 GB

RTX 5090 desktop limitations

575 W board power — budget for a 1000 W+ PSU and the heat it puts into the room.
Models larger than 32 GB must offload to system RAM, which costs roughly an order of magnitude in speed.

Recommended setup

Ollama or llama.cpp (CUDA/ROCm) — vLLM if you need concurrent serving

How these numbers are calculated

Memory need = quantized weights + KV cache at 8,192 tokens (f16) + 0.8 GB runtime overhead.
32 GB of VRAM on the NVIDIA GeForce RTX 5090 at 1792 GB/s.
64 GB of system RAM available for CPU offload when a model exceeds VRAM.
Throughput is a memory-bandwidth roofline estimate, not a measurement. It is labelled "estimated" everywhere it appears.
KV cache is computed from this model's published attention configuration.

FAQ