Can I Run Command R Family on RTX 4090 Desktop (24 GB VRAM, 64 GB RAM)?

Newer alternative. Cohere has not released a newer model in this library. If you are choosing today, Mistral Small 4 is a comparable current option. View Mistral Small 4 →

Written by Jakub Rusinowski · Last updated April 4, 2024

Yes, but it is tight

Yes, but it is tight — Command R (35B) at Q2_K needs about 23 GB of the 24 GB usable on RTX 4090 Desktop (24 GB VRAM, 64 GB RAM), leaving only ~1 GB before the runtime starts swapping. Expect ~40.7 tok/s (estimated), with room for about 8,192 tokens of context.

Confidence: medium · Recommended quantization: Q2_K · Estimated speed: ~40.7 tok/s

RTX 4090 Desktop (24 GB VRAM, 64 GB RAM) — what it gives a model

Usable memory for models24 GB
Memory bandwidth1008 GB/s
Form factorDesktop
Operating systemWindows or Linux
Memory upgradeableYes

Command R Family on RTX 4090 Desktop (24 GB VRAM, 64 GB RAM): memory by quantization

QuantMemory neededFits 24 GB?Max contextEst. speedDownload
F1681.5 GB✗ No70 GB
Q8_048.7 GB✗ No37.2 GB
Q6_K40.2 GB✗ No28.7 GB
Q5_K_M36.3 GB✗ No24.8 GB
Q4_K_M32.7 GB✗ No21.1 GB
Q3_K_M26.5 GB✗ No14.9 GB
Q2_K23 GB✓ Yes8K~40.7 tok/s11.5 GB

Which Command R Family sizes fit

VariantNeeds at Q4_K_MFits?Est. speed
Command R+ (104B)65.7 GB✗ Too large
Command R (35B)32.7 GB✗ Too large

What to watch out for

RTX 4090 desktop limitations

Recommended setup

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

How these numbers are calculated

FAQ

Can I run Command R Family on RTX 4090 Desktop (24 GB VRAM, 64 GB RAM)?

Yes, but it is tight — Command R (35B) at Q2_K needs about 23 GB of the 24 GB usable on RTX 4090 Desktop (24 GB VRAM, 64 GB RAM), leaving only ~1 GB before the runtime starts swapping. Expect ~40.7 tok/s (estimated), with room for about 8,192 tokens of context.

Which quantization of Command R Family should I use on RTX 4090 Desktop (24 GB VRAM, 64 GB RAM)?

Q2_K — it needs about 23 GB of the 24 GB available, downloads as roughly 11.5 GB, and runs at an estimated 40.7 tokens/sec with up to 8K of context.

What limits Command R Family on RTX 4090 Desktop (24 GB VRAM, 64 GB RAM)?

Nothing binding — the model fits with headroom and generates at a usable speed on this hardware.

Which runtime should I use?

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

Other Computers

Other Models on RTX 4090 Desktop (24 GB VRAM, 64 GB RAM)

Command R Family on GPUs

What This Model Is Good At

Model & Tools

← Can I Run It? | Command R Family model page | Check your hardware