VOOZH about

URL: https://willitrunai.com/models/command-r-plus-104b

โ‡ฑ Command R+ 104B VRAM Requirements โ€” GPU Compatibility


๐Ÿ‘ Cohere
Cohere

Command R+ 104B

Current
๐Ÿ‘ huggingface
HuggingFace๐Ÿ‘ ollama
Ollama
77Downloads286LikesApr 2024Released131K tokensContextCC-BY-NC-4.0License43 BasicQuality

Command R+ 104B (104B parameters) requires approximately 68.4 GB of VRAM with Q4_K_M quantization. For the best balance of quality and speed, we recommend hardware with at least 79 GB of VRAM.

Get started

โ€” copy & paste to run locally

Copy-paste commands to run Command R+ 104B on your machine.

Run

ollama run command-r-plus

Quick specs

Parameters104B
Architecturedense
Context131K tokens
Modalitytext
Min RAM40.6 GB
Rec. RAM63.4 GB (Q4_K_M)
LicenseCC-BY-NC-4.0
FamilyCommand
โœ“ Chatโœ“ Reasoningโœ“ RAG

About this model

Command R+ is Cohere's most capable open-weight model for enterprise RAG workloads. Offers superior long-context reasoning, multi-step tool use, and grounded generation with citations across 10 languages.

Related models

Your hardware

Detecting...

Quick picks

Best budgetB
MacBook Pro M3 Max 128GB~$2,499 โ€” 4 tok/s
๐Ÿ‘ NVIDIA
Best overallA
NVIDIA GH200 96GB~$30,000 โ€” 56 tok/s

Best hardware

Top picks for Command R+ 104B

NVIDIA GH200 96GBA
96 GB
NVIDIA H20 96GBA
96 GB
AMD Instinct MI300A 128GBA
128 GB
NVIDIA H200 141GBA
141 GB
NVIDIA H200 PCIe 141GBA
141 GB

Run this model

Command R+ 104B on NVIDIA GH200 96GBCommand R+ 104B on NVIDIA H20 96GBCommand R+ 104B on AMD Instinct MI300A 128GB

Quantization options

VRAM estimates by quant level

No hardware detected โ€” fit column shows raw VRAM estimates

QuantBitsVRAMQualityFit
Q2_K
2
40.6 GB
Lowโ€”
Q3_K_S
3
51.0 GB
Lowโ€”
NVFP4
4
58.2 GB
Mediumโ€”
Q4_K_M
4
63.4 GB
Mediumโ€”
Q5_K_M
5
74.9 GB
Highโ€”
Q6_K
6
85.3 GB
Highโ€”
Q8_0
8
111.3 GB
Very Highโ€”
F16
16
213.2 GB
Maximumโ€”

Quality benchmarks

Command R+ 104B benchmark scores

Benchmark verified

Coding

SWE-bench Verifiedโ€”
HumanEval+74.4%
Aider Polyglotโ€”
LiveCodeBenchโ€”

Reasoning

MMLU-Pro54.7%
GPQA Diamond13.4%
MATH-50046.0%
ARC Challengeโ€”

General

Chatbot Arenaโ€”
IFEval60.0%

Source: official ยท 2024-04-04

Hardware compatibility

Fit estimates across all hardware

Open calculator

Computing compatibility...

Memory breakdown

Reference: RTX 2060 6GB

Weights63.4 GB
KV Cache3.4 GB
Runtime0.9 GB
Headroom0.6 GB

Frequently asked questions

FAQ โ€” Command R+ 104B

See also

Quantization GuideScoring MethodologyVRAM Calculator