LLM Reference

Command Models by Cohere

CohereCC-BY-NC-4.0Open weights
10 models2023–2025Up to 256k ctxFrom $0.3/1M input

Details

ResearcherCohere
Commercial useCommercial use: non-commercial
Models10
Released2023–2025
Max context256k

Capabilities

Multimodal1 of 10 models
Function Calling5 of 10 models
Tool Use5 of 10 models
Structured Outputs6 of 10 models

About

Command is a family of 10 AI models by Cohere, released between 2023 and 2025.

Current Variants

Use-when guidance is based on each model's tracked capabilities, context window, release date, and replacement status.

8 in view2 retired

Use when the workload needs multilingual and 8k context.

2025-08multilingual8k context

Use when the workload needs reasoning, 256k context, and tool use.

2025-08reasoning256k contexttool use

Use when the workload needs vision, 128k context, and tool use.

2025-07vision128k contexttool use

Use when the workload needs chat, 256k context, and tool use.

2025-03chat256k contexttool use

Use when the workload needs 131k context, tool use, and function calling.

2024-08131k contexttool usefunction calling

Use when the workload needs 131k context, tool use, and function calling.

2024-08131k contexttool usefunction calling

Use when the workload needs 128k context and structured outputs.

2024-03128k contextstructured outputs

Use when the workload needs 128k context and structured outputs.

2024-03128k contextstructured outputs

Release Timeline

6 release groups
2025-08
2 current
Command A Reasoning (08-2025)
reasoning256k contexttool use
Current
Command A Translate (08-2025)
multilingual8k context
Current
2025-07
1 current
Command A Vision (07-2025)
vision128k contexttool use
Current
2025-03
1 current
Command A (03-2025)
chat256k contexttool use
Current
2024-08
2 current
Command R 08-2024
131k contexttool usefunction calling
Current
Command R+ 08-2024
131k contexttool usefunction calling
Current
2024-03
2 current
Cohere Command R
128k contextstructured outputs
Current
Cohere Command R Plus
128k contextstructured outputs
Current
2023-11
2 retired
Command
4k contextstructured outputs
Archived
Command Light
4k contextstructured outputs
Archived

Specifications(10 models)

Command model specifications comparison
ModelReleasedContextMultimodalFn CallingTool UseStructured Outputs
Command A Translate (08-2025)2025-088kNoNoNoNo
Command A Reasoning (08-2025)2025-08256kNoYesYesNo
Command A Vision (07-2025)2025-07128kYesYesYesNo
Command A (03-2025)2025-03256kNoYesYesNo
Command R+ 08-20242024-08131kNoYesYesYes
Command R 08-20242024-08131kNoYesYesYes
Cohere Command R2024-03128kNoNoNoYes
Cohere Command R Plus2024-03128kNoNoNoYes

Pricing

Command model pricing by provider
ModelProviderInput / 1MOutput / 1MType
Cohere Command RAWS Bedrock$0.5$1.5Serverless
Cohere Command R PlusIBM watsonx$1.2$3.6Serverless
Command R 08-2024Microsoft Foundry$1.5$2Serverless
Command A (03-2025)Microsoft Foundry$2.5$10Serverless
Command R+ 08-2024Microsoft Foundry$2.5$10Serverless
Cohere Command R PlusAWS Bedrock$3$15Serverless

Frequently Asked Questions

What is Command used for?
Command is used for multilingual, reasoning, and vision. The family description and listed model capabilities point to those workloads as the best fit.
How does Command compare to Cohere Transcribe?
Command by Cohere is strongest where you need multilingual, while Cohere Transcribe by Cohere is the closest related family to check for speech recognition. Command has 10 listed variants and reaches up to 256k context, so compare the specs and pricing tables before choosing a production model.
Which Command model should I use?
For the lowest listed input price, start with Command Light through Cohere API at $0.3/1M input tokens. For the most capable/latest local choice, evaluate Command R+ 08-2024 with 131k context and tool use, function calling, and structured outputs.