Mistral AI Le Chat Update July 2025
Mistral AI has introduced a series of updates to Le Chat designed to enhance research capabilities, multimodal interaction, and organizational workflow. These updates include the introduction of a Deep Research agent, a new voice-in model called Voxtral, and the integration of the Magistral reasoning model for multilingual tasks.
Deep Research Capabilities
Le Chat now features a Deep Research mode (currently in preview) that functions as a coordinated research assistant. This tool-augmented agent can plan research tasks, clarify user needs, search for information, and synthesize findings into structured, reference-backed reports.
Voice Interaction with Voxtral
Le Chat now supports a vocal mode that allows users to interact with the assistant via speech instead of typing. This functionality is powered by the new Voxtral model, which is designed for natural, low-latency speech recognition to support real-time brainstorming and transcription.
Multilingual Reasoning via Magistral
The "Think mode" in Le Chat is powered by Magistral, a reasoning model that enables the assistant to provide thoughtful answers to complex questions across multiple languages. Magistral supports native multilingual reasoning, allowing users to draft documents or clarify concepts in various languages and perform code-switching between languages within a single sentence.
Project-Based Organization
Le Chat has introduced Projects, which allow users to group related conversations into focused, context-rich spaces. Each project maintains its own default Library and remembers specific tool and setting configurations. Users can upload files directly into a project and integrate content from their Libraries to keep conversations and documents centralized for long-term workstreams.
Advanced Image Editing
In partnership with Black Forest Labs, Le Chat now offers advanced image editing capabilities. Users can modify generated images using simple prompts—such as removing objects or changing locations—while the model preserves characters and specific details. This allows for consistent visual elements across a series of images.
Sources
- OriginalLe Chat dives deep.
Related
- Dispatch
- Dispatch
- Dispatch
- Dispatch
- Dispatch