Back to models
Chat Models

Cohere

Command A

Best for Research

Text-focused Cohere model for grounded answers and tool workflows, with 256K context and 8K output.

RAG256K context
Open AppBrowse All Models

At a glance

Know the model before you prompt.

Cohere API specifications
Context window
256,000 tokens
Maximum output
8,000 tokens
Inputs → output
Text → Text
Knowledge cutoff
June 1, 2024

API model ID: command-a-03-2025

Capabilities & boundaries

What it supports. Where the limits are.

Tool support requires the appropriate API integration; a supported tool is not automatically active in every chat.

Supported API features and tools

  • A 111B-parameter model for text-based enterprise tasks
  • A 256,000-token context and 8,000-token maximum output
  • Grounded document answers and tool-oriented workflows
  • Support for 23 languages, subject to task-specific evaluation

Before you choose

  • This is the text-only Command A ID, not A Vision or A+; do not assume image input.
  • Do not attach Command A Reasoning's thinking controls to this different model ID.
  • Grounded generation can still misread a source or cite the wrong passage. Verify the cited evidence, not just the presence of a citation.
  • Context capacity is not a guarantee of complete recall. Keep source IDs, evaluate omissions and verify citations against the underlying documents.
  • A generated plan or tool call does not authorize or execute an external action. Use a configured integration, validate results and keep consequential changes behind human approval.

Reasoning behavior

The reviewed model-specific sources do not establish hybrid thinking controls or an effort-level list for this ID. Do not copy Command A Reasoning's thinking settings into other Command models. State the goal, evidence and answer format, then verify the result.

  • Use command-a-03-2025 for the model covered here. The website slug is a navigation label, not necessarily the API model ID.
  • Provide source documents or configured retrieval results; a prompt alone does not grant access to company data.
  • Give tools narrow schemas and explicit approval rules. Record returned tool results before generating a final answer.
  • Keep retrieval quality and model quality as separate evaluation dimensions. Missing evidence cannot be repaired by a confident answer.

Put it to work

Start with a more useful prompt.

Original examples from EZ Ai Assist. Adapt these to your task and the features available in your workspace.

Workflow 01

Answer a policy question with evidence

Make unsupported answers easy to spot.

Answer the employee question using only the policy excerpts below. Cite the document ID and section for each substantive claim. If policies conflict, explain the conflict without choosing an unstated precedence rule. Mark any missing information and finish with the question a policy owner must resolve. Do not invent policy or infer approval.

Workflow 02

Prepare a tool-assisted service plan

Keep read operations and write actions distinct.

Given this service request and the available tool descriptions, propose a sequence of read-only checks. State the evidence needed before each next step and identify any action requiring customer or operator approval. Use only the listed tool capabilities. Do not claim that a proposed tool call has run or that a customer record has changed.

Workflow 03

Review multilingual support consistency

Compare meaning while preserving policy boundaries.

Compare these support replies in their original languages against the supplied policy and terminology list. Identify differences in promises, eligibility and required next steps. Cite the exact passage for each issue and propose a correction for review. Flag ambiguous translation choices for a native speaker instead of silently changing meaning.

Developer reference

Cohere API pricing

These are Cohere API reference prices, not EZ Ai Assist subscription prices.

View EZ Ai Assist plans
Standard processing · USD per 1,000,000 tokens
Token typePrice
Input$2.50
Output$10.00
  • Rates come from the named model's current Cohere card. Other hosts, private deployments and commercial agreements can have different terms.
  • Input and output are billed separately. No model-specific cached-input rate is asserted here; absent cache pricing does not mean cached tokens are free.
  • Trial-key limits are separate from production usage. Check current key limits and actual billed usage before scaling a workflow.

Common questions

A few things worth knowing.

Which Command A ID does this page cover?

command-a-03-2025. A+, A Reasoning, A Vision and A Translate have different IDs, capabilities and limits; their features should not be combined into this record.

Can Command A inspect an image?

The model covered here is text-only. Use a documented vision model for image input, and confirm the application supports that input path.

Does 256K context mean 256K output?

No. The model card lists 256,000 context but only 8,000 maximum output tokens. Organize long inputs and request a bounded answer rather than assuming the output can mirror the entire context.

Does a citation prove an answer is correct?

No. It makes checking easier, but the cited passage may be incomplete or misinterpreted. Verify the passage, its date and whether it actually supports the claim.

Can it see current information automatically?

No. Model knowledge and a long context are not a live search service. Provide current evidence or connect an authorized retrieval workflow where supported, and verify the sources before acting on an answer.

How should I test it on my own work?

Use representative examples, explicit pass criteria and difficult counterexamples. Score factual support, omitted requirements and invalid outputs as well as latency and cost. Keep a human escalation path for uncertain or consequential results.

Are these API capabilities included in my EZ Ai Assist plan?

This guide describes the provider API, not subscription entitlements or a promise that every setting is exposed in the app. Check your workspace for model access and supported inputs, and use the EZ Ai Assist pricing page for plan details.

Check the source

Official documentation

Specifications and API prices checked on . Example prompts and workflow advice are editorial guidance from EZ Ai Assist.

Same provider

More from Cohere

Cohere

Command A+

Multimodal Cohere MoE with hybrid reasoning, 48-language support and a 128K context window.

Cohere

Command A Reasoning

Hybrid text reasoning with optional thinking budgets, 256K context and 32K output for evidence-heavy workflows.

Cohere

Command A Vision

Text and image understanding for documents, tables and charts, with up to 20 images per request and no tool use.

Cohere

Command A Translate

Specialized text translation across 23 languages, with separate 8K input and 8K output limits.