Toronto-based AI startup Cohere has released Rerank 3, a model designed to enhance enterprise search and retrieval augmented generation (RAG) systems. The model is claimed to be compatible with any database or search index, and can also be incorporated into any existing application with native search capabilities.

Additionally, the model has a 4k context length for better search results in longer documents, the ability to examine diverse and semi-structured data like emails, JSON files, invoices, programming codes, multilingual capabilities across 100+ languages, improved latency, and lower total cost of ownership.

Furthermore, the company claims that Rerank 3 integrates with RAG systems and enhances performance by providing high-precision semantic reranking, thus ensuring that the generative models process only the relevant results. 

Analyst QuickTake: The company has been releasing models continuously amidst its plan to raise substantial capital, ranging from USD 500 million to USD 1 billion. This month, the company launched Command R+ for business applications. Last month, Cohere AI released Command R for enterprise usage and the previous month launched a new open-source model, Aya .