IBM today announced the availability of Mistral Large on watsonx.ai, as part of our commitment to make high quality, performant foundation models widely accessible to enterprises – and deployable on their own terms. As IBM’s enterprise studio for AI developers, watsonx.ai provides access to a diverse selection of open-source and commercial models that they can customize and integrate with business solutions and applications.

This newest addition of Mistral Large, which will be initially available as SaaS, expands the collection of models accessible on watsonx.ai, including the IBM Granite family of foundation models – comprised of open code models, language models, and scientific models such as time series and geospatial –that are driving impact for our customers. With the availability of Mistral AI’s commercial model on our platform, we are further extending our open, multi-model strategy, empowering enterprises to innovate, adapt and scale.

Building on IBM’s commitment to advancing business innovation responsibly, we will also provide our customers with protection for Mistral Large via a capped intellectual property indemnity – making it the first third-party foundation model to which IBM extends indemnification.

IBM and Mistral AI are continuing to foster innovation and collaboration by providing enterprises with access to a mix of open-source and commercial models, and smaller, super-efficient models on watsonx.ai that prioritize data control. This approach is crucial for enhancing the quality of enterprise AI solutions.

Additionally, with Mistral Large now on the watsonx platform, enterprises can benefit from additional enterprise-ready products, including data store, prompt lab, model tuning, production monitoring and governance capabilities. Watsonx customers are not constrained by platform lock-in and are empowered to deploy models that are available on watsonx.ai in their environment of choice – whether on premise or a public cloud provider. In the fast-paced world of AI, flexibility is crucial for enterprises to adapt quickly and avoid sunk investments in infrastructure and development.

Watsonx customers can now benefit from Mistral AI’s most powerful model that has been optimized to handle complex enterprise tasks that require reasoning and multilingual capabilities. For example:

• RAG Specialization –supporting longer chat interactions and large document processing.
• Function Calling – allowing connections to external tools such as user defined functions or APIs. It enables ease building applications catering to specific use cases and practical problems.
• Coding Proficiency – generating, reviewing, and commenting on code, with the capability to output results in JSON format, a functionality that enables developers to interact with the model more naturally to extract information in a structured format.
• Multilingual Capability – offering, by design,  performance in French, German, Spanish, and Italian, in addition to English. Support is also provided for dozens of other languages.
• Responsible Al – it comes with built-in guardrail features with an additional layer available.

The combination of IBM, Mistral AI, and the watsonx platform provide an option for organizations operating in Europe to realize the benefits of foundation models while empowering them with governance capabilities, reaffirming IBM’s commitment to building responsible AI.

For more information about how you can access Mistral Large or the broader IBM and third-party model library available on watsonx.ai, visit: https://www.ibm.com/products/watsonx-ai/foundation-models