Databricks, led by Ali Ghodsi, has partnered with the Paris-based startup Mistral to integrate Mistral’s large language models (LLMs) into Databricks’ data intelligence platform. Databricks’ investment in Mistral, combined with bringing select LLMs to its platform, amplifies opportunities for enterprise users in generative AI applications without compromising security protocols. This move signifies Databricks as a prominent distribution partner for Mistral, joining recent collaborations with industry heavyweights such as Snowflake and Microsoft.
The partnership with Mistral introduces two text-generation models, Mistral 7B and Mixtral 8x7B, onto the Databricks Data Intelligence Platform. Mistral’s models showcase impressive capabilities, where Mixtral 8x7B notably outperforms competitors like Meta’s Llama 2 70B and OpenAI’s GPT-3.5 across various benchmark tests. Users can now access and customize these models on the Databricks platform, empowering them to build end-to-end generative AI applications efficiently.
While this collaboration benefits Databricks’ customers, it does not currently encompass Mistral Large, the high-performance model adjacent to GPT-4. Despite this, Mistral continues to expand its partnerships, such as securing a significant investment from Microsoft and forming collaborations with industry players like IBM, Perplexity, and Amazon.
With Mistral strategically broadening its reach through engagements with key tech companies, its trajectory in the AI landscape remains fascinating to observe. This development, along with Databricks’ integration of Mistral’s models, signifies a significant step forward in the evolution and adoption of advanced AI solutions across industries.