Shots Fired at ChatGPT as Mistral Adds Image Generation and Web Search

Platform Mistral adds a suite of new features to its chatbot, Le Chat, designed to make it all the all-in-one AI solution.

Mistral has launched a host of updates including adding image generation and web search to its chatbot, Le Chat.

The French startup also unveiled an open source multimodal AI model called Pixtral Large, which can handle up to 30 high-resolution images per input and is driving the new tools, which Le Chat now includes.

The company’s offerings – especially the free Le Chat – have gathered fans worldwide and impressed Microsoft enough to give $16 million for the company to incorporate its LLM Mistral Large into its cloud computing platform Azure.

Mistral’s Le Chat Gets Chattier

Le Chat is free to use but you need to set up a Google, Mistral or Microsoft account to access it. Users can then experiment with the new features by enabling them as beta features, said Sophia Yang, the company’s Head of Developer Relations.

Heading up the new features is a real-time web search option, which now includes source citations so users can see exactly where information is coming from. ChatGPT does offer citations for where it’s pulling information from in its own chatbot search, but this data is limited on the free versions.

 

About Tech.co Video Thumbnail Showing Lead Writer Conor Cawley Smiling Next to Tech.co LogoThis just in! View
the top business tech deals for 2024 👨‍💻
See the list button

Next up is Advanced Document and Image Analysis. Le Chat is able to summarize information from PDFs, extracting key data or presenting in a different way. These tools already exist on other platforms, but having an all-in-one solution built into Le Chat is a nice touch. Yang showed how this works with a receipt uploaded using OCR. Le Chat was able to read the information, and then split the bill as requested, including a tip.

Ideation and Image Innovation

With “creative ideation” as the driver, the new tools also include an interactive “canvas”, which Yang says can be used for everything from creating documents and presentations to code. It allows for collaboration direct from the chatbot interface.

This is a direct competitor to OpenAI’s Canvas sidebar, which was launched only last month. “This early beta introduces a new way of working together—not just through conversation, but by creating and refining ideas side by side”, says OpenAI. The French team may also have Anthropic’s Artifacts option in their sights.

Mistral has teamed up with image model startup Black Forest Labs to offer an image generation option too. The German company has offered up its Flux Pro LLM to power the new tool. In another sabre-rattle at OpenAI, this tool goes up against DALL-E 3’s integration into ChatGPT.

Automation Options

Last up are a host of what Mistral call “Task Agents”, which can tackle repetitive tasks like processing invoices or providing meeting summaries. Mistral explains that these agents can be customized.

Users already had the option of creating agents either through La Plateforme Agent builder or the company’s Agent API; but now this has been honed for Le Chat.

Mistral adds: “Agents can be published and shared with your teammates. You can ideate recursively, and build increasingly powerful agents that help you get your work done faster.”

Is Mistral a Serious Rival to OpenAI?

With these new tools, the Mistral team is all about user experience, and CEO and co-founder Arthur Mensch wrote on his account on the social network X that the company’s new LLM was designed with just this in mind. He said: “At Mistral, we’ve grown aware that to create the best AI experience, one needs to co-design models and product interfaces. Pixtral was trained with high-impact front-end applications in mind and is a good example of that.”

Pixtral Large follows hot on the heels of the Mistral Large 2 model, which arrived in the summer and the multimodal model, Pixtral 12-B. Mistral makes their models available for free download but the license stipulates that usages must be non-commercial, research-focused applications.

Its positioning as a top AI offering from Europe is definitely gathering the platform followers, especially as US tech leaders become more embroiled in US politics. It may behind OpenAI in many respects – including in the enterprise space – but it is iterating fast and obviously has huge ambitions.

Did you find this article helpful? Click on one of the following buttons
We're so happy you liked! Get more delivered to your inbox just like it.

We're sorry this article didn't help you today – we welcome feedback, so if there's any way you feel we could improve our content, please email us at contact@tech.co

Written by:
Katie has been a journalist for more than twenty years. At 18 years old, she started her career at the world's oldest photography magazine before joining the launch team at Wired magazine as News Editor. After a spell in Hong Kong writing for Cathay Pacific's inflight magazine about the Asian startup scene, she is now back in the UK. Writing from Sussex, she covers everything from nature restoration to data science for a beautiful array of magazines and websites.
Explore More See all news
Back to top
close Building a Website? We've tested and rated Wix as the best website builder you can choose – try it yourself for free Try Wix today