Invest in Cohere

Cohere provides large language models and RAG capabilities tailored to meet the needs of enterprise use cases. The Command family of models includes Command R+ and Command R.

Command R+ is Cohere’s latest large language model, designed for exceptional performance in conversational interactions and long-context tasks. It empowers companies to seamlessly transition from proof of concept to full-scale production.

Command R+ is trained on an extensive and diverse multilingual dataset, enabling it to handle a broad range of text-generation tasks. Additionally, it is specifically optimized to excel in key business applications, addressing some of the most critical use cases with precision and efficiency.

Command R is a large language model tailored for conversational interactions and long-context tasks. Positioned in the "scalable" category, it strikes a balance between high performance and strong accuracy, empowering companies to transition confidently from proof of concept to production.

Embed models are designed to generate embeddings from text or classify it based on various parameters. These embeddings can estimate semantic similarity between texts, predict the most likely sentence to follow another, or categorize user feedback. Paired with the Classify endpoint, embeddings enable versatile applications for classification and analysis tasks across a range of use cases.

Cohere's Rerank endpoint enhances search quality by integrating a semantic-based reranking system into traditional keyword-driven search. By contextualizing the meaning of user queries beyond simple keyword relevance, Rerank delivers significantly higher-quality results compared to embedding-based search alone. With support for documents and queries in over 100 languages, it requires just a single line of code to implement, making it an effortless upgrade for any application.

Fine-tuning is a powerful method for tailoring LLMs like Cohere Command R to specific domains and tasks, delivering substantial performance gains over the base model.

Fine-tuned versions of Command R models have shown over 20% improvement across diverse enterprise applications in industries such as financial services, technology, retail, healthcare, and legal.

Due to its compact size, a fine-tuned Command R models can be deployed more efficiently, outperforming much larger models in its class.

Cohere’s revenue model is structured with flexible, usage-based pricing across its advanced generative and retrieval AI models, designed to support diverse enterprise applications.

For generative AI, customers can choose from models like Command R+, optimized for large-scale tasks, and Command R, ideal for retrieval-augmented generation and other long-context applications. Retrieval options include Rerank 3 for enhancing search relevance and Embed 3 for semantic retrieval, both of which allow companies to leverage AI for search, data insights, and customer interactions.

Cohere is uniquely positioned as a key beneficiary of Canada's strategic investments in AI innovation. Earlier in 2024, the Canadian government announced over $2 billion in subsidies, including computing power, to support the country’s AI startups, with Cohere among those receiving backing. This national support underscores Canada’s commitment to fostering homegrown model developers and building a distinct AI ecosystem, independent of U.S.-centric AI companies and regulatory challenges in the EU. With additional funding from PSP Investments, one of Canada’s largest pension funds, Cohere is well-equipped to scale its operations and contribute significantly to Canada’s leadership in the global AI landscape.

According to The Information, in July 2024, Cohere was generating approximately $35 million in annualized revenue around the time of its last funding round.

According to The Information, in March 2024, Cohere informed investors that its sales pipeline, representing the total value of potential contracts that could close by the end of 2024, exceeds $300 million.

In March 2024, Cohere announced it partnered with Accenture to deliver its advanced Command and Embed models along with cutting-edge Retrieval Augmented Generation capabilities, empowering organizations to scale generative AI adoption across their enterprises. Lan Guan, Accenture's Chief AI Officer, highlighted the partnership in her blog, stating: “With Accenture’s $3 billion investment in AI, we’ve joined forces with Cohere, a leading enterprise AI platform, to accelerate how we help companies integrate AI into their operations at scale, with a strong focus on privacy and security.”

In April 2024, Cohere announced that its latest Command R model family is now available on Amazon Bedrock. This milestone builds on Cohere’s partnership with AWS to deliver top-tier AI solutions, empowering enterprise customers to drive significant productivity and efficiency improvements in their operations.

In July 2024, Cohere announced the availability of its Rerank 3 model on Microsoft Azure AI Studio. Leveraging Cohere’s enterprise-grade language models on Azure AI’s powerful infrastructure allows businesses to seamlessly, securely, and reliably integrate advanced semantic search technology into their applications.

In September 2024, Cohere partnered with Japanese consulting firm Nomura Research Institute (NRI) to announce the NRI Financial AI Platform. This platform, powered by Cohere’s secure, enterprise-grade AI technology, aims to enhance productivity and optimize operations for global financial institutions. The platform is set to launch in the first half of 2025.

In December 2024, Cohere partnered with CoreWeave to build a multibillion-dollar data center in Canada, with financial help from the Canadian government. The Canadian government The government will support Cohere for as much as $170 million as part of its Canadian Sovereign AI Compute Strategy.

In December 2024, TechCrunch reported that Palantir is a partner of Cohere, whose models are already being used by several unnamed Palantir customers. During Palantir’s inaugural developer conference, DevCon1, held in November 2024, Cohere engineer and former Palantir employee Billy Trend stated that Cohere is “already deploying to Palantir customers.”