LICENSEWARE
Publisher

Cohere

Cohere licenses its Command, Embed, Rerank, Parse and Transcribe models and its North and Compass applications through a pay-as-you-go API, single-tenant Model Vault instances, hyperscaler marketplaces and private deployments, each governed by a different agreement.

Start Here

Overview articleCohere licensingCohere licenses its Command, Embed, Rerank, Parse and Transcribe models and its North and Compass applications through a pay-as-you-go API, single-tenant Model Vault instances, hyperscaler marketplaces and private deployments, each governed by a different agreement.Cited · Edited 2026-10-13 · 14 references

Guides

Licensing Data

From the Cohere catalog: how usage is counted, the rules that apply, the programs licences are bought through and the SKUs. Backed by 90 verbatim excerpts from 31 Cohere documents.

Generative AI Software 6

Cohere API, Cohere Aya, Cohere Command, Cohere Command A, Cohere Command R, Cohere North

Machine Learning Software 4

Cohere Embed, Cohere Model Vault, Cohere Parse, Cohere Rerank

Transcription Software 1

Cohere Transcribe

Product History

All acquisitions
  1. Cohere acquired Reliant AIAnnounced 2026-05-19 · Closed 2026-05-19

    Cohere announced that it had acquired Reliant AI, a biopharma AI company with operations in Montreal and Berlin, to expand its sovereign enterprise AI offering in healthcare and life sciences.

    Legacy productBecame
    Reliant AI research workbenchMerged into North for Pharma · Cohere said Reliant's technology would be integrated to build North for Pharma
    Licensing impact
    • Cohere said it would bring Reliant AI's research team, proprietary biomedical datasets and domain-optimised technology into its sovereign AI platform and accelerate North for Pharma, an agentic system for biopharma teams.[1]
    • Reliant AI's CEO Karl Moritz Hermann joins Cohere as VP of AI Verticalizations and co-founder Marc Bellemare as VP of Modelling.[1]
    Sources and full product history

Litigation

All cases
  1. Advance Local Media v. CohereUS District Court for the Southern District of New York · 2026
    Pending
    Lessons learned
    • A court let claims proceed that a model's retrieval-augmented outputs, not only its training, can reproduce or closely paraphrase protected articles.
    • Marketing a model for a use, such as summarising the latest news, was treated at the pleading stage as possible evidence of inducement.
    • Investigator-obtained outputs were accepted at the pleading stage as evidence of typical third-party use, despite a terms-of-service objection.
Esc