LLM APIs · B tier
Cohere, B tier
73/100
Retrieval stacks that need embeddings and a reranker more than a celebrity chat model.
Enterprise API focused on retrieval, rerank, and command models. A specialist, not the default chat brain.
Best for
Retrieval stacks that need embeddings and a reranker more than a celebrity chat model.
What it is
Cohere sells an API aimed at enterprise retrieval: embed, rerank, and its Command generation models. It is a strong specialist for search stacks and a weaker default than OpenAI or Anthropic for a general product assistant.
Score breakdown
Weights: capability 40, ease 25, value 20, momentum 15. Momentum is the editor’s judgment. How the score works.
Pricing
Freemium
Starting price: check site
A trial key and usage-based production billing. Enterprise contracts are separate. Check the pricing page.
Dates
Released 2019 (year precision).
Added to StackPick 14 Oct 2026.
Alternatives
Questions
What is Cohere best for?
Retrieval stacks that need embeddings and a reranker more than a celebrity chat model.
How much does Cohere cost?
Freemium. Starting price: check site. A trial key and usage-based production billing. Enterprise contracts are separate. Check the pricing page.
What are alternatives to Cohere?
Mistral, OpenAI API, Amazon Bedrock are the alternatives we file next to it. Open a comparison when one exists.
How do model leaderboards relate to Cohere?
Cohere is scored here as a product. OpenRouter rankings (https://openrouter.ai/rankings) inform momentum from token usage. Arena (https://arena.ai) informs capability from blind votes on the models it serves. Figures stay on those sites.