Today, we’re introducing Cohere Embed V5 Pro and Embed V5 Fast in Microsoft Foundry, bringing stronger multimodal retrieval to enterprise search and AI applications. With Fast and Pro options, support for more than 100 languages, and flexible vector sizes, Embed V5 gives developers more control over retrieval quality and efficiency.
Find relevant information across text and images
Enterprise knowledge spans documents, wikis, reports, and images. Embed V5 helps applications find relevant information across these sources and languages, supplying richer context to search experiences and retrieval-augmented generation (RAG) workflows.
According to Cohere, Embed V5 surpasses Embed V4 across multimodal, multilingual, and domain-specific search and retrieval benchmarks. For AI agents, better retrieval can mean more useful context, better answers, and fewer unnecessary tokens spent processing irrelevant information.
Choose Fast or Pro to match your workload
Embed V5 introduces two options. Fast is designed for latency-sensitive workloads, such as interactive search and responsive agent experiences. Pro prioritizes maximum retrieval quality for applications where finding the most relevant information is critical.
Both share a single embedding space, supporting interchangeable use and giving teams flexibility as their application requirements evolve.
Scale retrieval with more control over cost
A broader range of Matryoshka embedding dimensions lets developers choose smaller vectors to reduce storage requirements and retrieval latency while targeting the quality their application needs. High-throughput batch processing supports embedding large content collections for indexing at scale.
Beyond search and RAG, Embed V5 also supports classification, clustering, recommendations, and similarity matching, helping teams put enterprise data to work across more applications.
Get started in Microsoft Foundry
Explore Cohere Embed V5 for your next search or RAG application. Evaluate Fast and Pro on your own enterprise content to find the right balance of relevance, latency, and cost.
Explore Cohere Embed V5 in Microsoft Foundry.
Cohere-Embed-V5-Pro | Model Catalog | Microsoft Foundry