Command R
Cohere's scalable enterprise LLM optimized for retrieval-augmented generation and tool use with a 128K context window.
Cohere 
Amazon Bedrock Enterprise RAG and tool use with 128K context
Command R is an instruction-following conversational model developed by Cohere, designed for enterprise language tasks with a focus on reliability and scalability. It is available through Amazon Bedrock and carries a knowledge cutoff of March 2024. The model is purpose-built for retrieval-augmented generation (RAG) and tool use, making it well-suited for workflows that require grounding responses in external data sources or integrating with external APIs and functions.
One of Command R's defining characteristics is its 128,000-token context window, which allows it to process long documents, extended multi-turn conversations, and complex inputs in a single pass. It also supports multilingual tasks and is tagged for low-latency performance, making it a practical choice for organizations building scalable AI applications where response speed and contextual accuracy matter. It is best suited for enterprise use cases such as document analysis, agentic pipelines, and knowledge-grounded question answering.
What Command R supports
Retrieval-Augmented Generation
Grounds model responses in external knowledge sources by retrieving and citing relevant documents, reducing hallucinations in enterprise workflows.
Tool Use & Function Calling
Enables agentic workflows by allowing the model to call external tools and APIs, supporting multi-step task execution.
Long Context Understanding
Processes up to 128,000 tokens in a single pass, enabling analysis of long documents and extended multi-turn conversations.
Multilingual Support
Handles tasks across multiple languages, making it suitable for globally distributed enterprise applications.
Low Latency Responses
Optimized for speed in production environments, supporting real-time or near-real-time application requirements.
Instruction Following
Follows detailed natural language instructions reliably, supporting structured enterprise use cases such as summarization, classification, and Q&A.
Ready to build with Command R?
Get Started FreeBenchmark scores
Scores represent accuracy — the percentage of questions answered correctly on each test.
| Benchmark | What it tests | Score |
|---|---|---|
| MMLU-Pro | Expert knowledge across 14 academic disciplines | 33.8% |
| GPQA Diamond | PhD-level science questions (biology, physics, chemistry) | 28.4% |
| MATH-500 | Undergraduate and competition-level math problems | 16.4% |
| AIME 2024 | American math olympiad problems | 0.7% |
| LiveCodeBench | Real-world coding tasks from recent competitions | 4.8% |
| HLE | Questions that challenge frontier models across many domains | 4.8% |
| SciCode | Scientific research coding and numerical methods | 6.2% |
Common questions about Command R
What is the context window size for Command R?
Command R supports a context window of 128,000 tokens, allowing it to process long documents and extended conversations in a single pass.
What is the knowledge cutoff date for Command R?
Command R has a training knowledge cutoff of March 2024, as noted in the model metadata.
How is Command R priced on Amazon Bedrock?
Command R is available through Amazon Bedrock, and pricing is determined by AWS. You can find current pricing details on the Amazon Bedrock Pricing page.
What is Command R best suited for?
Command R is purpose-built for retrieval-augmented generation (RAG) and tool use, making it well-suited for enterprise workflows that require grounding responses in external knowledge sources or building agentic pipelines.
Does Command R support multiple languages?
Yes, Command R is tagged as multilingual and is designed to handle tasks across multiple languages, supporting globally distributed enterprise applications.
What people think about Command R
Community discussions reflect that Command R was widely used as a daily driver model among developers, particularly for its RAG capabilities and enterprise focus. Users in the LocalLLaMA subreddit noted it had a strong following roughly two years ago before newer models shifted attention.
More recent threads focus less on Command R specifically and more on the broader Cohere ecosystem, including the company's $6.8 billion valuation and ongoing enterprise positioning. Some community members have raised questions about where Command R fits as Cohere's model lineup has evolved.
What ever happened to Cohere’s Command-R and Command-A series of models? R was a lot of folks’ daily driver model like 2 years ago.
GPT -OSS is heavily trained on benchmark. scored rank 34 on simplebench worse than grok 2
AI startup Cohere valued at $6.8 billion in latest fundraising, hires Meta exec
Parameters & options
Explore similar models
Cohere Command R+
Cohere's most powerful Command model, optimized for complex enterprise workloads requiring advanced reasoning, RAG, and multi-step tool use.
OpenAI GPT‑5.2 Pro
The most capable model series yet for professional knowledge work.
Google Gemini 3.1 Pro
Google's frontier reasoning model delivering enhanced software engineering, agentic reliability, and multimodal intelligence across a 1M-token context window.
DeepSeek DeepSeek-R1
Reasoning LLM from Chinese AI company DeepSeek utilizing Chain of Thought.
Amazon Amazon Nova Lite
Very low cost multimodal model that is lightning fast for processing image, video, and text inputs.
Nvidia Nemotron 3 Nano 30B
NVIDIA's compact open-weight hybrid MoE model with 30B total / 3.5B active parameters, delivering strong reasoning and coding performance up to 1M context.
Start building with Command R
No API keys required. Create AI-powered workflows with Command R in minutes — free.