In summary
- Le Chat de Mistral integrated Flux Pro for image generation, surpassing DALL-E 3 in precision and detail.
- Pixtral Large scored 93.3% on DocVQA, beating Google’s GPT-4 and Gemini 1.5 Pro in key visual tests.
- Mistral offers free features such as web search and document analysis, matching paid tools from OpenAI.
Mistral AI, the French artificial intelligence company founded by former Google and Meta employees, has improved its free multimodal AI platform—and is coming close to matching, and sometimes even surpassing, the paid offerings of OpenAI or Anthropic.
The Paris-based startup’s AI toolset now features web search and image generation capabilities along with a new visual model that outperforms larger rivals in key benchmarks.
The company has promoted an update to analyze documents, an updated Large Language Model (LLM), and Codestral, an LLM focused on programming tasks.
But the highlight is its use of Flux Pro to generate images, making it the multimodal chatbot with the best image generation capabilities currently available. A new Canvas interface allows for direct content editing, positioning Le Chat—the name Mistral chose for its chatbot interface—as an enterprise-ready workspace.
“We are not pursuing AGI at all costs; our mission is instead to put frontier AI in your hands,” Mistral said in a statement, distancing itself from its competitors’ broader artificial general intelligence ambitions that have raised concerns among experts.
The company was valued at $2 billion in its December funding round.
Pixtral Large, a 124 billion parameter visual model, outperformed GPT-4 and Google’s Gemini 1.5 Pro with 69.4% accuracy in MathVista and 93.3% in DocVQA. Handles up to 30 high-resolution images within a 128,000-token context window.
Flux Pro, meanwhile, outperforms all other imagers in terms of quality and inference speeds, as Decrypt previously reported.
For its part, ChatGPT uses DALL-E, which is primitive by comparison. So the integration of Flux into Le Chat is a welcome move.
These features have been introduced as “free beta offerings”, giving users access to the latest generative AI tools at no cost. Mistral did not say when it intended to start charging for its tools.
ChatGPT vs. Mistral LeChat: How They Compare
We tested the free versions of the Mistral and ChatGPT chatbots. There were many similarities—mainly in writing style—but the differences were enough to draw some distinctions between the two.
ChatGPT’s GPT-4o remains the crown jewel of language models, displaying unparalleled nuance in complex reasoning, creative writing, and technical analysis.
His ability to understand context and maintain long, coherent conversations has defined the industry standard. During extensive testing, GPT-4o has demonstrated remarkable capability in tasks ranging from literary analysis to advanced programming challenges.
Le Chat Large 2.5, while impressive, operates a level below GPT-4o in sophisticated reasoning tasks. However, it excels in practical applications, offering agile responses and robust performance in daily tasks. While the model has multilingual capabilities, it tends to be less creative than GPT-4o.
Web Search Integration
Le Chat offers a free web search feature – users can access current information with direct citation links, similar to how you would use a traditional search engine but with AI synthesis. The integration feels natural and unobtrusive, making fact-checking and investigation workflows remarkably efficient. Free access to this feature represents a significant value proposition.
ChatGPT’s SearchGPT, while more sophisticated in its approach, is behind a paywall. It takes a unique angle on web integration, often providing more contextual analysis (to the point of even boggling some answers). The Chrome plugin offers deeper integration with OpenAI’s language model capabilities, although many users might find Le Chat’s direct approach more practical for everyday use.
Imaging Capabilities
The integration of Flux Pro into Le Chat is a huge plus in the imaging space. The system produces highly detailed images with great adherence to prompts, particularly excelling in artistic styles, precise human anatomy and complex scene composition. Recent benchmarks place Flux Pro at the forefront of imaging technology, with noticeably better performance in areas such as texture detail and spatial coherence.
ChatGPT’s DALL-E 3, while revolutionary at launch, is now showing its age against newer models. Although capable of producing usable images, it often struggles with complex prompts, detailed textures and maintaining consistent style – areas where Flux Pro demonstrates clear superiority. The gap is particularly noticeable in professional-grade imaging tasks.
Custom AI Agents
Le Chat’s democratic approach to AI agents represents a significant change in accessibility. Users can create, customize, and deploy specialized agents for tasks ranging from data analysis to creative writing for free.
ChatGPT, on the other hand, maintains a two-tier system. While free users can access pre-built GPTs, the ability to create custom ones remains a premium feature.
The quality of these GPTs is excellent, and the ability to configure them using natural language prompts is a plus; but the restriction on authoring tools and the price wall limits the potential for custom workflow optimization.
Visual Analysis Capabilities
There seems to be a tie here—at least in our first quick tests. GPT-4V and Pixtral Large demonstrate remarkably similar capabilities in real-world applications.
Both excel in diagram interpretation, image understanding, mathematical notation, and spatial awareness.
Verdict
While the benchmark numbers show slight variations for LLMs, Vision Models, and Imaging Models, the practical difference in everyday use cases is not that significant.
Both systems handle everything from creative writing to technical diagrams with impressive precision, making the choice between them more about ecosystem preference than capability differences.
The choice between these platforms increasingly depends on specific use cases:
ChatGPT is still superior for:
- academic writing
- Complex technical documentation
- Nuanced creative writing
- Advanced programming challenges
Le Chat offers best value for:
- Visual content creation
- Web research
- Document analysis
- Daily productivity tasks
- Workflow automation
The financial implications are the deciding point: Le Chat offers approximately 80-90% of the functionality of ChatGPT Plus free of charge, making it an increasingly attractive option for both individual users and small businesses with budget constraints.
Overall, Mistral offers a better value proposition today, and its models may serve users well enough to justify canceling a subscription—at least for the duration of the beta testing period.
Edited by Josh Quittner and Sebastian Sinclair
Generally Intelligent Newsletter
A weekly AI journey narrated by Gen, a generative AI model.
Crypto Keynote USA
For the Latest Crypto News, Follow ©KeynoteUSA on Twitter Or Google News.
