IonBin
All pages
Search
Model · Cohere

Cohere models (Command, Embed, Rerank)

Cohere builds models for enterprise retrieval and generation: the Command family for chat and agents, Embed and Rerank for search, plus parsing, transcription and multilingual models. It is available on its own platform and through major clouds, which matters if procurement wants AI inside an existing cloud contract.

Verified Oct 18, 2026 Re-verify: past review date 3 sources

What it is

Cohere is an enterprise-focused model vendor. Its docs group models into generation (Command), retrieval (Embed and Rerank), and supporting models for parsing documents, transcription and translation.

How it works

  1. Use a Command model to generate text or run tool-using agents. Command A+ is the newest and accepts text and images.

  2. Use Embed to turn documents into vectors for search, then Rerank to reorder the top results before the model answers. That is the core of RAG.

  3. Deploy through Cohere’s API, through a cloud marketplace your company already buys from, or on dedicated capacity through Model Vault.

Specs and pricing (checked 2026-10-18). Command A+ has a 128k context window and 64k output tokens. Trial API keys are free but rate-limited and can’t be used commercially. Production keys are pay-as-you-go. Model Vault rents dedicated retrieval models by the hour, from $3/hour for a small Embed instance.

Benefits

  • Retrieval models (Embed, Rerank) are first-class products, which helps internal search projects.

  • Cloud availability can fit an existing procurement route, but the specific model, region and contract still need review.

  • Embed v5.0 supports 128k inputs; chunk size still affects retrieval precision and should be tested.

Limits and risks

  • Build-it-yourself option. These API/model options require an application or integration; Cohere also sells the North workspace.

  • Version churn. Model names and versions change often. Pin versions in production.

  • Pricing takes work to find. We couldn’t find per-token prices for current Command models on the public pricing page. Get a quote before you budget.

Checked specs

Snapshot from official sources on Oct 18, 2026. Model lineups and prices change often; this page is re-verified every 30 days.

Newest Command modelCommand A+ (command-a-plus-05-2026): Cohere's first mixture-of-experts model; text and images; 128k context, 64k output [1]
Self-hosting footprint (Command A+)Fits on one B200 or two H100 GPUs, per Cohere [1]
Other Command modelsCommand A, Command A Reasoning, Vision, Translate; Command R7B, R+, R [1]
Retrieval modelsEmbed v5.0 (pro, fast; 128k context); Rerank v4.0 (pro, fast) [1]
API keys (checked 2026-10-18)Trial keys free but rate-limited and not for commercial use; Production keys pay-as-you-go, billed monthly or when $250 is outstanding [2]
Model Vault, dedicated hosting (checked 2026-10-18)Embed 5 Fast or Pro, Small: $3/hour or a listed $2,000 monthly rate; Medium: $5/hour or a listed $3,250 monthly rate; Rerank 4 Pro, Large: $10/hour or a listed $6,500 monthly rate [2]
Per-token prices for current Command modelsNot shown in the pricing page text we checked; ask Cohere or check the cloud marketplace listing [2]
PlatformsCohere, Bedrock, SageMaker, Azure AI Foundry, Oracle [1]

Best-fit stacks

Stack 05

Customer support

Answer faster from a help desk your agent can search, with help-center drafts a person approves.

Stack 08

Dev, web & IT

Design, code, host and ship sites and internal tools, with an AI coding agent doing most of the typing.

Agent compatibility

API accessYesOfficial docs ↗
Also offered viaAmazon Bedrock, Amazon SageMaker, Azure AI Foundry, Oracle (per Cohere docs)

Levels come from each tool's recorded, sourced compatibility. How the levels work.

Sources

  1. An overview of Cohere's models · Cohereofficial accessed Oct 17, 2026
  2. Pricing · Cohereofficial accessed Oct 17, 2026
  3. What is a vector database? · IBMreference accessed Oct 17, 2026

Practitioner views are restated in our words and labeled as views, not measured results. Product facts come from official sources dated above.

Suggest an edit

IonBin's wiki isn't user-edited. Corrections go to our Directory Curator, who checks them against primary sources before anything changes. Spot something wrong or out of date on “Cohere models (Command, Embed, Rerank)”? Email us with a link to a primary source.

Suggest an edit by email or write to contact@ionbin.com