Mistral AI adds Deep Research agent, voice mode with Voxtral model to Le Chat
Mistral AI has released a major update to Le Chat, adding a Deep Research agent that generates structured research reports, a new voice input model called Voxtral, and Projects for organizing conversations. The update also includes multilingual reasoning powered by Mistral's Magistral model.
Mistral AI adds Deep Research agent, voice mode with Voxtral model to Le Chat
Mistral AI has released a major update to Le Chat, its AI assistant platform, adding research capabilities, voice input, and new organizational tools.
Deep Research Agent
The new Deep Research mode (in preview) functions as a tool-augmented agent that plans research queries, searches for information, and synthesizes findings into structured, reference-backed reports. According to Mistral AI, the agent can "break down" complex questions, gather credible sources, and produce reports with citations.
Voxtral Voice Model
Le Chat now supports voice input through a new model called Voxtral. The voice-in model is designed for "natural, low-latency speech recognition," allowing users to interact with the assistant through speech rather than text. Mistral AI claims the system can keep pace with natural speaking speeds.
Magistral Reasoning and Multilingual Support
The update includes a "Think mode" powered by Magistral, Mistral's reasoning model. According to the company, the model provides "clear, thoughtful answers" for complex questions and supports multilingual reasoning, including code-switching between languages within a single conversation. The company cites examples like drafting proposals in Spanish or clarifying legal concepts in Japanese.
Projects and Image Editing
Le Chat now includes Projects, a feature for organizing related conversations into folders with context retention. Each project maintains its own library and remembers tool settings.
The platform has also added image editing capabilities through a partnership with Black Forest Labs. Users can edit generated images with text prompts like "remove the object" or "place me in another city." Mistral AI claims the system preserves character and object consistency across edits.
Availability
All features are available in Le Chat at chat.mistral.ai and through mobile apps on iOS and Android. Mistral AI states no credit card is required to access the features. Pricing for the new capabilities was not disclosed.
What this means
Mistral AI is expanding Le Chat beyond basic conversational AI into research and multimodal territory, directly competing with features in ChatGPT, Claude, and Perplexity. The addition of Voxtral puts Mistral in the voice assistant space alongside OpenAI's Advanced Voice Mode and Google's Gemini Live. The Deep Research agent positions Le Chat as a potential alternative to Perplexity's research-focused interface, though comparative performance benchmarks are not yet available. The key question is whether Mistral's models can match the quality of established players in these specific use cases.
Related Articles
Mistral Rebrands Le Chat as Vibe, Launches Agentic Work and Code Modes with VS Code Extension
Mistral has rebranded Le Chat as Vibe, launching new agentic capabilities for long-running work tasks and software development. The platform now includes Work Mode for enterprise knowledge search and document synthesis, Code Mode with GitHub integration and sandboxed execution, and a new VS Code extension. Pricing starts at $14.99/month for Pro and $24.99/user/month for Team plans.
Mistral Releases Voxtral TTS: 4B Parameter Text-to-Speech Model at $0.016 per 1k Characters
Mistral AI has released Voxtral TTS, a 4B parameter text-to-speech model supporting 9 languages including English, French, German, Spanish, Dutch, Portuguese, Italian, Hindi, and Arabic. The model achieves 70ms latency for typical inputs and can clone voices from as little as 3 seconds of audio, priced at $0.016 per 1,000 characters.
Mistral OCR 3 launches at $2 per 1,000 pages with 74% win rate over previous version
Mistral AI released Mistral OCR 3, a document extraction model priced at $2 per 1,000 pages ($1 with Batch API discount). The model achieves a 74% overall win rate over its predecessor on forms, scanned documents, complex tables, and handwriting according to internal benchmarks.
Mistral Releases Mistral 3 Family: 675B-Parameter Large 3 MoE and Three Edge Models Under Apache 2.0
Mistral has released Mistral 3, including Mistral Large 3—a sparse mixture-of-experts model with 41B active and 675B total parameters—and three Ministral 3 edge models (3B, 8B, 14B). All models are released under Apache 2.0 license with multimodal capabilities and are available today on multiple platforms.
Comments
Loading...