Cohere104B~64 GB VRAM at Q4_K_M

Command R+ (104B) — VRAM, speed & local setup

Written by Jakub Rusinowski · Last updated

Cohere's flagship model for complex enterprise RAG. Handles multi-hop retrieval, grounded citations, and complex tool orchestration. Outperforms GPT-4 Turbo on RAG benchmarks.

Command R+ (104B) needs about 64 GB of VRAM at Q4_K_M — quantized weights plus framework overhead, before any KV cache. On Apple Silicon that figure comes out of unified memory.

Logic94
Creative87
Coding88

VRAM and speed by quantization

Quoted against NVIDIA RTX 4090 (24 GB). Includes the KV cache at 8K context, so it reads higher than the headline figure.

QuantVRAMSpeed (est.)Fit
Q2_K
2.63 bpw
37.1 GB~3 tok/sOffload
Q3_K_M
3.41 bpw
47.3 GB~2 tok/sOffload
Q4_K_M
4.83 bpw
65.7 GB—Too big
Q5_K_M
5.67 bpw
76.7 GB—Too big
Q6_K
6.56 bpw
88.2 GB—Too big
Q8_0
8.50 bpw
113.4 GB—Too big
F16
16.00 bpw
210.9 GB—Too big

Black marker = usable memory on the NVIDIA RTX 4090 (24 GB). Estimates from the memory-bandwidth roofline on the methodology page. Command R+ (104B) VRAM calculator →

Get Command R+ (104B) running

The cheapest catalogued GPU that runs Command R+ (104B) is the NVIDIA DGX Spark (128 GB).

Buy this hardware NVIDIA DGX Spark (128GB) — 128 GB VRAM · 150 W board powerDeploy in the cloud now NVIDIA A100 80GB on RunPod — from $1.39/hr · rate checked 2026-07

or compare on Vast.ai from $0.77/hr (typical low · varies)

As an Amazon Associate we earn from qualifying purchases. Cloud GPU links are referral links — we may earn a commission at no extra cost to you.

Affiliate disclosure: Some links on this page are affiliate links — if you buy through them, LLM Configurator may earn a commission at no extra cost to you. As an Amazon Associate, LLM Configurator earns from qualifying purchases.
NVIDIA DGX Spark (128GB)
128 GB VRAM · 150 W board power
2026 prices are volatile — check the current listing.

How to run Command R+ (104B)

Install Ollama, then run:

ollama run command-r-plus
Weights on Hugging Face: CohereForAI/c4ai-command-r-plus ↗

Specifications

Curated — A hand-written entry from before this catalogue recorded its sources. The figures are long-standing but their provenance is not on file.

Parameters
104 Billion
Context window
128,000
Architecture
Dense
Provider
Cohere
Licence
CC-BY-NC
Practical deployment
Workstation (large unified memory or pro GPU)
Specified at
Q4_K_M
System RAM
128 GB
Record updated
2024-04-04
LicenceCC-BY-NC-4.0Research / non-commercial only

Research / non-commercial only — this licence does NOT permit shipping a commercial product.

Quality and use cases

Scores as published by the model’s authors or an independent evaluator — quality, not throughput, and not measured by us.

Best forenterprise ragagentcomplex retrieval

Can I run Command R+ (104B) on my GPU?

Other Command R Family sizes

Command R+ (104B) — frequently asked questions

How much VRAM does Command R+ (104B) need?

About 64 GB at Q4_K_M — quantized weights plus framework overhead, before any KV cache. The cache grows with context length and is added on top; the table above folds it in. Apple Silicon counts unified memory toward the same figure.

Does Command R+ (104B) run on an RTX 4090 (24 GB)?

No. Command R+ (104B) needs about 64 GB at Q4_K_M, more than a single RTX 4090's 24 GB. It needs a larger card, several GPUs, or Apple Silicon with enough unified memory — or it runs with part of the weights offloaded to system RAM, which is much slower.

How do I run Command R+ (104B) locally?

Install Ollama and run `ollama run command-r-plus`. That pulls the weights and starts a local OpenAI-compatible endpoint; after the download nothing leaves the machine.

What other sizes does Command R Family come in?

Command R (35B) (22 GB), Command R+ (104B) (64 GB). Every size shares the family's training and licence; the larger ones score higher and need proportionally more memory.