Learn About Cohere's AI Products
Create a free account to access in-depth lessons on each tool and model.
Start Learning FreeπAbout Cohere
Updated June 15, 2026Cohere is a Canadian AI company founded in 2019 by Aidan Gomez (co-author of the original "Attention Is All You Need" Transformer paper), Ivan Zhang, and Nick Frosst. The company develops large language models specifically optimized for enterprise use cases, with a focus on retrieval-augmented generation (RAG), text analysis, and multilingual capabilities.
Cohere's model portfolio includes Command R+ (a powerful model optimized for RAG and tool use with 128K context), Embed (state-of-the-art text embeddings for search and classification in 100+ languages), and Rerank (a model that improves search result relevance). Cohere has also moved into open models with North Mini Code β its first agentic coding model and first open-weights release, a 30 billion-parameter mixture-of-experts model under the permissive Apache 2.0 license, a strategic step beyond its closed Command lineup. The company differentiates from competitors by offering flexible deployment options β Cohere models can run on any major cloud provider, on-premises, or even on-device, making it particularly attractive to enterprises with data sovereignty requirements.
Cohere has raised over $1 billion in funding and serves enterprise customers across financial services, technology, and healthcare. The company's Coral conversational AI assistant and its focus on enterprise data privacy (with SOC 2, HIPAA eligibility, and GDPR compliance) position it as a leading choice for organizations that need production-grade language AI without sending sensitive data to consumer-focused AI providers.
π οΈProducts & Tools (6)
Cohere enterprise agentic AI platform (launched summer 2025) deployed entirely behind customer firewalls. Built on Command A (111B, 256K context). Customers include RBC, Dell, and LG CNS. No data leaves your infrastructure β critical for regulated industries.
Cohere's first agentic coding model and first open-weights release β a 30 billion-parameter mixture-of-experts design (3 billion active) with a 256K context, Apache 2.0 weights on Hugging Face.
Cohere open-weight 3.35B-parameter multilingual models for 70+ languages. Regional variants: Global, Earth (Africa/Middle East), Fire (South Asia), Water (Asia-Pacific/Europe). Designed for laptops and edge devices. Apache 2.0 license. Complements Command A for resource-constrained deployment.
Cohere's flagship enterprise model. 111B parameters, 256K context, runs on just 2 GPUs. Matches GPT-4o on enterprise agentic tasks. Includes Translate (23 languages SOTA) and Reasoning variants.
Cohere's open-source multilingual model supporting 23 languages. Now expanded with Aya Vision (multimodal) and Tiny Aya (3.35B, 70+ languages for edge/laptop deployment).
Cohere's enterprise AI assistant built on Command A (111B, 256K context). Purpose-built for RAG in regulated industries. North platform deploys behind customer firewalls. $7B valuation, IPO expected 2026.
