Mistral AI Unveiled Major Updates to Le Chat

French startup Mistral AI has launched major updates to generative AI assistant Le Chat, which is now in beta version for free. Make Le Chat your creative collaborator, train it to carry out particular duties, and use it as a copilot when coding. In other words, use it to accomplish tasks.

What’s New?

Le Chat is getting additional powers from Mistral AI (in beta). Here is a brief synopsis:

• Citations in a web search
• State-of-the-art document and image understanding, powered by the new multimodal Pixtral Large
• Canvas for ideation, in-line editing, and export
• Fully integrated offering, from models to outputs
• Speculative editing-powered quicker reactions
• Image generation, powered by Black Forest Labs Flux Pro

Also Read: What is Mistral AI’s New Model Ministraux?

Let’s look at the features in detail now.

Web search

Students and professionals use Le Chat for study, research, and work, accounting for a sizable portion of its usage. Mistral AI added an internet search to Le Chat’s training knowledge to better support these use cases.

Here’s an illustration.

You have a strong instrument to elevate your job with state-of-the-art language models that have access to the most recent information and events.

Also Read: Mistral AI Unveiled Mistral-Small Instruct-2409

Canvas

When you need to move past chats and into brainstorming, a new interface called Canvas appears in your Le Chat window. It allows you to work together with Mistral LLMs on common outputs, facilitating knowledge work ideation and 0-1 development.

With the help of LLMs’ potent reasoning powers, Canvas allows you to edit, change, or transform content directly, eliminating the need for bidirectional discussions.

Here is an example of a video.

Documents, presentations, code, mockups, and more may all be made with Canvas. As you can see, you can preview your designs, version your drafts, and change their contents in place without having to generate responses.

Any team can generate endless creative ideas with Canvas.

Also Read: What is Mistral AI La Plateforme? How to use it to Create AI Agents?

Quick and precise comprehension of documents

Large, complicated PDF documents and photos can now be processed by Le Chat for analysis and summary. To obtain insights on anything, including graphs, tables, diagrams, text, formulas, equations, and pretty much the entire document, just attach papers to your chat box.

Together with an experimental model (more on that in a few weeks), the new Pixtral Large, a state-of-the-art multimodal model, enables these sophisticated capabilities. It is noticeably more potent than our previous vision model, Pixtral 12B.

The following example demonstrates how Mistral AI can extract information, summarize, and comprehend semantics from the well-known 1935 paper on quantum entanglement by Einstein, Podolsy, and Rosen.

Also read: Mistral Announces the Release of the AI Agent Building Platform 

Creation of images

Mistral AI is bringing their top picture generation models to Le Chat in collaboration with their pals at Black Forest Labs. Now, you may create excellent photos for your job directly from Le Chat.

Le Chat Agents will automate tedious tasks.

You can codify these workflows as agents if you frequently reuse prompts or if you’ve noticed specific creative work patterns that you’d like to see automated. Processing invoices, generating TL/DRs of lengthy meeting minutes, and scanning receipts for expense reporting are common use cases.

You can share agents with your peers and publish them. Recursive ideation allows you to create progressively more potent agents that speed up your task.

What is the total cost of this?

All of these features are now available as free beta versions.

Also Read: Google Cloud Partners with Mistral AI to Boost Vertex AI with Codestral Code Generation