Cohere
Verified Oct 16, 2026 ActiveEnterprise LLM family (Command, Embed, Rerank, Parse, Transcribe) sold API-first with VPC, on-prem and dedicated hosting options
Listed on EffectHub ·
Pricing
- Pricing model
- Freemium
Trial keys are free, rate-limited and barred from commercial use; production keys are pay-as-you-go per token, billed monthly or whenever the balance reaches $250. The pricing page lists per-token rates only for legacy models (Command $1/$2, Command R+ 08-2024 $2.50/$10 per 1M in/out); current Command, Embed, Rerank, Parse and Transcribe rates are not shown there. Model Vault dedicated instances run $3 to $10 an hour ($2,000 to $6,500 a month); North, Compass, private deployments and model customization are contact-sales.
Use tool ↗About Cohere
Cohere builds enterprise language models and sells them API-first with an emphasis on private deployment. The model line now spans Command for generation and agents, Embed 5 and Rerank 4 for retrieval, Parse for document parsing and Transcribe for speech, while North 2 (agent platform) and Compass (search) are quote-only. Trial API keys are free but rate-limited and non-commercial; production keys bill pay-as-you-go per token, though the pricing page now shows per-token rates only for legacy models kept for existing customers. Dedicated Model Vault instances are the one fixed price published, from $3 an hour or $2,000 a month for an Embed 5 instance. Founded in Toronto in 2019, it suits regulated enterprises that need VPC, on-prem or sovereign hosting more than hobbyists.
- Platforms
- Web · API
- API
- Yes
- Last verified
- Oct 16, 2026
Verified sources
- Trial keys free, rate-limited and non-commercial; production keys pay-as-you-go billed monthly or at a $250 balance; per-token rates shown only for legacy models (Command $1/$2, Command R+ 08-2024 $2.50/$10 per 1M); Model Vault dedicated instances $3 to $10 an hour, $2,000 to $6,500 a month; North, Compass and private deployments contact-sales (15 October 2026)cohere.com ↗
- Product line lists North 2, Compass, Command, Embed 5, Rerank, Parse, Transcribe and North Small Translate, with Model Vault and private deployments in isolated VPCs or on-prem; homepage announces an Aleph Alpha sovereign-AI agreement (15 October 2026)cohere.com ↗
- Founded in 2019 in Toronto by Aidan Gomez, Nick Frosst and Ivan Zhang; offices in Toronto, New York, London, San Francisco, Montreal, Paris, Seoul and Berlin; nearly $1 billion raised across four rounds 2021 to 2024 (15 October 2026)cohere.com ↗
- Terms of Use are an agreement with Cohere Inc. governed by the laws of Ontario and the federal laws of Canada (15 October 2026)cohere.com ↗
Is Cohere legit?
We don’t label tools safe or unsafe. Below are the records we checked ourselves, each with the date we read it.
- Domain registered
- Mar 7, 2000 (26 years ago)Registry record (RDAP), registrar Tucows Domains Inc., looked up Sep 6, 2026
- TLS certificate
- Issued by Let's Encrypt, valid until Nov 21, 2026Seen in our own connection to the site on Oct 16, 2026
- Last checked
- Oct 16, 2026The site answered
- On the market since
- 2019The vendor's own material, read by us
- Operated by
- Cohere Inc., CanadaOur research, card verified Oct 15, 2026
Pros & cons of Cohere
Cohere is the pragmatic choice when retrieval quality and data sovereignty drive the decision — Embed and Rerank remain reference points, and few rivals match its VPC and on-prem menu. The catch is that its generative models trail the frontier labs on raw capability, and the newest platform products push you into sales conversations rather than self-serve.
Pros
- Embed and Rerank are widely used building blocks for enterprise search
- Deployment menu spans SaaS API, VPC, on-prem and managed dedicated instances
- Clear per-token public pricing for API models
- Free rate-limited keys make evaluation easy
Cons
- Generation quality sits behind top-tier frontier models
- Flagship platforms (North, Compass) have no public pricing
- Trial keys prohibit commercial use
Best for
- Enterprises in finance, healthcare and government needing private or sovereign deployment
- RAG pipelines that lean on high-quality embeddings and reranking
- Multilingual products using Command's 49-language coverage
Not for
- Hobbyists wanting a consumer chatbot subscription
- Teams chasing the strongest general-purpose frontier model
Watch out: The pricing page still headlines legacy Command models; confirm current-generation model pricing and availability with the docs before committing.
Alternatives to Cohere
Cohere head-to-head
Side-by-side fact sheets against the competitors our research or their makers name — pricing, platforms, API and both verdicts on one page.