Le Chat dives deep

In this product update, Mistral AI announces enhancements to Le Chat, with the direction of making the AI assistant better suited to long-form thinking, research, cross-lingual work, and multimodal content organization. The five core new features in this release are: a Deep Research preview mode, Voice mode, native multilingual reasoning, Projects, and advanced image editing within Le Chat. Deep Research turns Le Chat into a working mode that is more like a research assistant: when facing a complex question, it can first plan the task, clarify needs, search for information, and then compile a structured report with sources, with the goal of reducing the user's burden of breaking down the problem, verifying, and organizing information on their own. Voice mode lets users converse with Le Chat without typing, suitable for brainstorming while walking, quick queries while out, or for meeting transcription; this feature is powered by Mistral's Voxtral speech model, emphasizing natural, low-latency speech recognition. On multilingual reasoning, Le Chat's Think mode is powered by Magistral, emphasizing that it can handle more complex problems in the user's familiar language and can switch languages within a sentence, which is of more practical value to multinational teams, researchers, creators, and non-English-speaking workers. Projects is a knowledge and workflow management feature that lets you place related conversations into contextual folders, retaining a default Library, enabled tools, and settings, and also lets you upload documents or reference Library content, suitable for moving plans, product feature design, or long-term projects. The image editing feature is in collaboration with Black Forest Labs, where users can use text instructions to request removing objects, changing scenes, and so on, while trying to preserve the consistency of characters, objects, and design elements. Overall, this is not a single model release, but a product upgrade that integrates Le Chat from a chat tool toward research, voice, multilingual support, project management, and image workflows.