Skip to content
Updated minutes ago
Cohere

Command R+

Cohere · dense · 104B parameters · 131,072 context

Quality
68.0

Parameters

104B

Context Window

128K tokens

Architecture

Dense

Best GPU

B200 SXM

Cheapest API

$2.00/M

Quality Score

68/100

Intelligence Brief

Command R+ is a 104B parameter DENSE model from Cohere, featuring Multi-Head Attention (MHA) with 64 layers and 12,288 hidden dimensions. With a 131,072 token context window, it supports tools, structured output, code, multilingual. On standardized benchmarks, it achieves MMLU 80, HumanEval 50, GSM8K 88. The most cost-effective API deployment is via together at $2.00/M output tokens. For self-hosted inference, B200 SXM delivers optimal throughput at $4261/month.

Architecture Details

TypeDENSE
Total Parameters104B
Active Parameters104B
Layers64
Hidden Dimension12,288
Attention Heads96
KV Heads96
Head Dimension128
Vocab Size256,000

Memory Requirements

BF16 Weights

208.0 GB

FP8 Weights

104.0 GB

INT4 Weights

52.0 GB

KV-Cache per Token3145728 bytes
Activation Estimate3.00 GB

GPU Compatibility Matrix

Command R+ is compatible with 21% of GPU configurations across 41 GPUs at 3 precision levels.

BF16 (Full)
FP8 (Half)
INT4 (Quarter)
Blackwell(7 GPUs)
B200 NVL (pair)360GB
B300288GB
B100 SXM192GB
GB200 NVL72 (per GPU)192GB
Hopper(7 GPUs)
H100 NVL 94GB (per GPU pair)188GB
H200 SXM141GB
H2096GB
GH20096GB
Ada Lovelace(11 GPUs)
L40S48GB
L4048GB
RTX 6000 Ada48GB
L2048GB
Ampere(16 GPUs)
A100 80GB SXM80GB
A100 80GB PCIe80GB
A1664GB
RTX A600048GB
Legend:No fitVery tightTightModerateGoodExcellent

GPU Recommendations

B200 SXMoptimal

FP8 · 1 GPU · tensorrt-llm

100/100

score

Throughput

280.0 tok/s

Latency (ITL)

3.6ms

Est. TTFT

1ms

Cost/Month

$4261

Cost/M Tokens

$5.79

Use this config →
B100 SXMoptimal

FP8 · 1 GPU · tensorrt-llm

100/100

score

Throughput

280.0 tok/s

Latency (ITL)

3.6ms

Est. TTFT

1ms

Cost/Month

$4271

Cost/M Tokens

$5.80

Use this config →
GB200 NVL72 (per GPU)optimal

FP8 · 1 GPU · tensorrt-llm

100/100

score

Throughput

280.0 tok/s

Latency (ITL)

3.6ms

Est. TTFT

1ms

Cost/Month

$6169

Cost/M Tokens

$8.38

Use this config →

Deployment Options

API

API Deployment

together

$2.00/M

output tokens

Self-Hosted

Single GPU

B200 SXM

$4261/mo

Min VRAM: 104 GB

Scale

Multi-GPU

H100 SXM x2

280.0 tok/s

TP· $3587/mo

API Pricing Comparison

ProviderInput $/MOutput $/MBadges
together$2.00$2.00
Cheapest
cohere$2.50$10.00

Cost Analysis

ProviderInput $/MOutput $/M~Monthly Cost
togetherBest Value$2.00$2.00$20
cohere$2.50$10.00$63

Cost per 1,000 Requests

Short (500 tok)

$1.40

via together

Medium (2K tok)

$5.60

via together

Long (8K tok)

$20.00

via together

Performance Estimates

Throughput by GPU

B200 SXM
280.0 tok/s
B100 SXM
280.0 tok/s
GB200 NVL72 (per GPU)
280.0 tok/s

VRAM Breakdown (B200 SXM, FP8)

Weights
KV
Act
Weights 104.0 GBKV-Cache 25.8 GBActivations 24.0 GBOverhead 5.2 GB

Precision Impact

bf16

208.0 GB

weights/GPU

fp8

104.0 GB

weights/GPU

~280.0 tok/s

int4

52.0 GB

weights/GPU

Quality Benchmarks

Average
71th percentile across all models
MMLU
80.0
Average (59th pctile)
HumanEval
50.0
Below Average (47th pctile)
GSM8K
88.0
Average (62th pctile)
MT-Bench
83.0
Bottom 25% (0th pctile)

Capabilities

Features

Tool Use Vision Code Math Reasoning Multilingual Structured Output

Supported Frameworks

vllmsglangtgitensorrt-llm

Supported Precisions

BF16 (default)FP8INT4

Where to Deploy Command R+

Similar Models

Command R

35B params · dense

Quality: 68

from $0.50/M

Cheaper, Smaller modelCompare →
Cheaper, Smaller modelCompare →

Yi-Large

102.6B params · moe

Quality: 74

from $3.00/M

Smaller context, Higher qualityCompare →

Inflection 3

100B params · dense

Quality: 74

from $15.00/M

Smaller context, Higher quality, More expensiveCompare →
Smaller context, Lower qualityCompare →

Frequently Asked Questions

How much VRAM does Command R+ need for inference?

Command R+ requires approximately 208.0 GB of VRAM at BF16 precision, 104.0 GB at FP8, or 52.0 GB at INT4 quantization. Additional VRAM is needed for KV-cache (3145728 bytes per token) and activations (~3.00 GB).

What is the best GPU for Command R+?

The top recommended GPU for Command R+ is the B200 SXM using FP8 precision. It achieves approximately 280.0 tokens/sec at an estimated cost of $4261/month ($5.79/M tokens). Score: 100/100.

How much does Command R+ inference cost?

Command R+ API inference starts from $2.00/M input tokens and $2.00/M output tokens. Self-hosted inference costs depend on your GPU configuration — use our ROI calculator for a detailed breakdown.