French AI lab Mistral has taken a significant step forward with its Le Chat chatbot, unveiling a suite of new features that bring it on par with industry giants like OpenAI and Google. The update introduces a "deep research" mode, native multilingual reasoning, and advanced image editing capabilities, enhancing Le Chat's functionality and versatility.

Following the release of Voxtral, Mistral's open-source AI audio model capable of multilingual reasoning and transcription, the company is now integrating Voxtral into Le Chat, expanding its capabilities further. Mistral's head of product, Elisa Salamanca, explains that the deep research mode transforms Le Chat into a coordinated research assistant. It can plan, clarify user needs, search, and synthesize data from a variety of sources on the web.

This feature is particularly relevant for both consumer and enterprise use cases. For consumers, Le Chat can research travel and provide an exhaustive analysis of the best travel plans. For enterprise work, it can conduct exhaustive research, making it a valuable tool in various sectors.

Mistral is focusing on integrating Le Chat and its productivity suite into enterprise ecosystems with this release. The company handles data connectors differently than competitors, addressing the needs of customers with sensitive data who prefer not to use cloud services or to use them on their own premises with virtual private clouds. Salamanca cites customers in banking, defense, and government as examples.

To cater to these needs, Mistral's Le Chat and productivity suite connect to enterprise data on-premises. This allows companies to use Le Chat's deep reasoning and other capabilities to analyze their own internal data without having to upload anything to the cloud. This is a significant differentiator against cloud-native LLM platforms like OpenAI, which is Azure-hosted, and Gemini, which is Google Cloud-based.

Mistral is also working on making Microsoft Excel, Google Docs, and other office suites work seamlessly with Le Chat's capabilities. The company is building these connectors internally, as it believes this will be key to using Le Chat as a productivity enabler in the business context.

In addition to deep reasoning, Le Chat has received other updates. Mistral's new reasoning AI model, Magistral, previously only available in English, now supports multilingual reasoning in languages like French, Spanish, Japanese, and more. Le Chat can also code-switch between languages mid-sentence.

The latest update also includes the addition of Projects, which helps users stay organized by grouping chats, documents, and ideas into focused spaces. Each project can have its own default library and remember which tools and settings a user has enabled. Use cases include planning a move, designing a new product feature, or keeping work-related projects on track.

Finally, Le Chat is getting improved image-editing capabilities, allowing users to create and edit images with prompts like "remove the object" or "place me in another city."

These new features put Mistral in a more competitive range with the AI industry's frontrunners, positioning Le Chat as not just a model demo but also a full-stack contender.