llmgauge

run it locally

Run Command R 35B on RTX 3090

35B parameters on a 24 GB card. It fits at Q3_K_M — here's the VRAM math at every quantization, the speed to expect, and the quality you trade.

verdict · Command R 35B on RTX 3090
Fits at Q3_K_M
needs 18.1 GB at Q3_K_M · 22.8 GB usable · ~53 tokens/sec
Q4_K_M won't fit here — run it in the cloud
Smallest GPU that fits at Q4_K_M: A100 40GB
Rent A100 40GB by the hour
VRAM needed and fit for Command R 35B on RTX 3090 by quantization.
quantization vram needed fits 24gb? tokens/sec quality
FP16 (full) 84.0 GB ✗ no — Reference quality — no quantization loss.
Q8_0 44.5 GB ✗ no — Near-lossless; rarely worth the extra space over Q6.
Q6_K 34.4 GB ✗ no — Virtually indistinguishable from full precision.
Q5_K_M 29.0 GB ✗ no — Minor loss; an excellent quality-vs-size balance.
Q4_K_M 23.5 GB ✗ no — Small but measurable loss; the popular default.
Q3_K_Mrecommended 18.1 GB ✓ yes ~53 Noticeable degradation; only when you're tight on VRAM.

Weights = params × bytes/weight, +20% for KV cache & runtime; usable VRAM is 95% of nameplate. Tokens/sec is a bandwidth ceiling (936 GB/s) — real throughput is lower with long context. Try other combinations →

Running Command R 35B on a RTX 3090

At Q3_K_M, Command R 35B's weights take about 18 GB, inside the RTX 3090's 23 GB of usable memory. Noticeable degradation; only when you're tight on VRAM. Generation speed is bound by memory bandwidth — the GPU reads the whole model once per token — so expect on the order of 53 tokens/sec before context overhead. To run it at the popular Q4_K_M default you'd want a A100 40GB or larger — or rent one by the hour rather than buy.

Quantization is the lever

Each step down in precision shrinks the model: FP16 needs about 84 GB, Q4_K_M about 24 GB — a 72% reduction for a small, usually acceptable quality cost. Q5_K_M and Q6_K are near-lossless if you have the headroom; drop to Q3 only when you're genuinely out of VRAM. The quantization guide covers the tradeoffs in detail.

Frequently asked questions

Can a RTX 3090 run Command R 35B?

Yes — Command R 35B fits on a RTX 3090 at Q3_K_M (18 GB of the 23 GB usable), with roughly 53 tokens/sec. Higher-precision quants need more VRAM; the table above shows each option.

How much VRAM does Command R 35B need?

Command R 35B is a 35B-parameter model. At FP16 that's about 84 GB; at Q4_K_M (the popular default) about 24 GB, including ~20% for the KV cache and runtime. Quantization is the main lever — see the per-quant table above.

Other combinations

Command R 35B on other GPUs: RTX 3060 12GB, RTX 4070 Ti, RTX 4080, RTX 4090, NVIDIA L4, RTX A6000

Other models on the RTX 3090: Gemma 2 9B, Phi-3 Medium 14B, Gemma 2 27B, Qwen2.5 32B, Mixtral 8x7B, Llama 3.3 70B

Related