Mistral Launches Free Chat Interface With Web Search, Canvas Editor, and Image Generation

Mistral Launches Free Chat Interface With Web Search, Canvas Editor, and Image Generation

Mistral AI has released major updates to le Chat, its free conversational AI assistant, adding web search with citations, a collaborative canvas editor, document understanding, and image generation capabilities. All features are available free in beta as of November 18, 2024.

New Capabilities

Le Chat now includes web search functionality that augments the model's training knowledge with real-time internet data and provides citations. The canvas interface allows users to create and edit documents, presentations, and code directly within the chat window, with in-place modifications and version control.

Document and image understanding is powered by Pixtral Large, Mistral's new multimodal model that the company claims is "significantly more powerful" than its previous vision model, Pixtral 12B. The system can process large PDFs, analyzing graphs, tables, diagrams, text, and equations.

Image generation capabilities come through a partnership with Black Forest Labs, using their Flux Pro model to generate images directly within le Chat.

Agent Automation

Le Chat supports custom agents that automate repetitive workflows. Users can create agents for tasks like receipt scanning, meeting summarization, and invoice processing. These agents can be shared with teammates and built recursively for more complex automation.

Competitive Positioning

Mistral positions le Chat against ChatGPT, Perplexity, and Claude. According to the company's comparison, le Chat offers free access to web search with citations, canvas editing, image generation, and image analysis—features that competitors either don't offer or charge for. ChatGPT requires paid access for canvas and agents, while Perplexity charges for image analysis and lacks canvas and image generation entirely.

Pricing Strategy

All features are currently free in beta. Mistral states it is "working towards premium tiers with higher service guarantees" but hasn't disclosed pricing or timelines. The company emphasizes its "frugal" approach to capital, claiming to deliver "frontier capabilities at affordable price points."

Technical Details

The system uses speculative editing for faster response times. Document understanding combines Pixtral Large with what Mistral describes as "an experimental model" that will be detailed "in a few weeks." The interface is available at chat.mistral.ai, with features rolling out gradually over the coming weeks.

What This Means

Mistral's aggressive free tier directly challenges OpenAI and Anthropic's pricing models, particularly for users who need web search and collaborative editing. The combination of Pixtral Large's multimodal capabilities with free image generation creates a genuinely differentiated offering, though the sustainability of keeping these features free remains unclear. The agent-building functionality could be particularly valuable for teams automating workflows, provided the underlying models prove reliable at scale.