Cohere releases powerful ‘Command-R’ language model for enterprise use

Join leaders in Boston on March 27 for an exclusive night of networking, insights, and conversation. Request an invite here.

Artificial intelligence startup Cohere announced the release of a major new language model called Command-R today, as the Toronto-based company finds itself in the midst of a heated fundraising round that could bring in as much as $1 billion in fresh capital.

Command-R represents a significant leap forward for Cohere’s technology, offering enhanced performance on key AI tasks like retrieval augmented generation (RAG) and tool use, longer context windows up to 128,000 tokens, and more affordable pricing.

“Command-R was designed to handle large-scale production workloads across the languages of global business,” said Cohere President & COO Martin Kon in an interview with VentureBeat. “We’ve optimized for RAG to combine accuracy and efficiency, which works even better with our Embed and Rerank models, helping enterprises move past the proof of concept stage.”

Cohere CEO Aidan Gomez said on Twitter that the new model is “smarter, longer context, cheaper” than the company’s previous Command model.

VB Event

The AI Impact Tour – Boston

We’re excited for the next stop on the AI Impact Tour in Boston on March 27th. This exclusive, invite-only event, in partnership with Microsoft, will feature discussions on best practices for data integrity in 2024 and beyond. Space is limited, so request an invite today.

Request an invite

The release comes at a pivotal time for Cohere, which is locked in an arms race with rival AI startups like OpenAI and Anthropic. Founded in 2019 by Gomez and fellow ex-Google researchers, Cohere has emerged as one of the leading AI companies focused on building powerful language models for enterprise use cases.

Targeting the enterprise market

While OpenAI has captured mainstream attention with the viral success of its ChatGPT chatbot, Cohere has taken a more targeted approach, working closely with business customers to tailor its models for their specific needs. This has allowed Cohere to operate more cost efficiently than competitors pursuing broad consumer applications.

“It is really important to build trust with enterprises and move beyond the proof of concept stage of AI into production,” explained Kon. “This is why Cohere focuses on privacy and data security, and also ensures that customers can access our models on all the major cloud providers to avoid vendor lock-in.”

The new Command-R model, which combines the company’s Embed and Rerank technology, tops the performance chart in end-to-end retrieval augmented generation (RAG) tasks, outpacing open source alternatives and leading commercial models like GPT-3.5-turbo. (Credit: Cohere)

Even so, developing state-of-the-art AI is immensely capital intensive. Cohere has already raised over $500 million to date, achieving a $2.2 billion valuation in its most recent funding round in June 2023. Now, the company is back at the negotiating table, with sources indicating it could raise anywhere from $500 million to $1 billion at an even higher valuation. 

Proving the business model

The high-stakes fundraising effort reflects the immense promise investors see in AI, but also the growing pressure on startups like Cohere to prove they can turn cutting-edge research into profitable businesses. As the generative AI market matures, companies will need to demonstrate not just impressive tech, but real-world customer adoption and revenue growth.

“Command-R should help our new and existing customers scale up quickly and get into large scale production,” said Kon. “Some of our current customers and partners include Oracle, Notion, Scale AI, Accenture and McKinsey.”

Kon pointed to Scale AI’s Gen AI Platform as an example of how Cohere’s models have driven tangible results. “One example is Scale AI’s Gen AI Platform who works with us to build a custom knowledge management application for their customer support team,” he said. “They used Cohere’s models to optimize TCO while maintaining high-level performance.”

Cohere appears to be making strides on the business front. In addition to its new Command-R model, the company recently opened a second headquarters in San Francisco to be closer to key customers and partners. It has also grown its workforce to over 250 employees.

Earlier this month, the company also announced a New York office that will serve as a hub for its leadership team to connect with existing partners in the city.

The coming months will be critical for Cohere as it looks to secure the war chest needed to compete against deep-pocketed rivals. But with the release of Command-R and a sharpening focus on the enterprise, Gomez and his team have positioned themselves as a startup to watch in the fast-moving world of artificial intelligence.

“There is a lot of noise with flashy models that are not practical to put into production,” said Kon. “It’s really important to look the Scalable category of AI models, like Command-R, that can deliver real results with the efficiency to handle heavy workloads.”

VentureBeat’s mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Discover our Briefings.

Source link

About The Author

Scroll to Top