Misconceptions about Mistral funding
Mistral AI reached a valuation of €21 billion, or roughly $24 billion, after closing a €3 billion Series D round on September 8, 2026. This amount exceeds the $6.2 billion valuation from June 2024. The company, founded in 2023 by Arthur Mensch, Guillaume Lample, and Timothee Lacroix, expands its European footprint through a new hub in Munich. Samsung Electronics led the recent funding round. Mistral provides EU-native data processing, which helps businesses in regulated sectors like finance or healthcare meet GDPR requirements without negotiating custom enterprise agreements with US providers. The expansion of Mistral’s infrastructure includes a $830 million investment in new data centers near Paris and in Sweden to support the company’s goal of bringing frontier AI into the hands of European users while maintaining jurisdictional sovereignty over sensitive workloads. This funding push aims to scale the company’s ability to compete with US giants. Mistral also received a €1.3 billion investment from ASML in September 2025. Earlier stages of growth included a €105 million seed round and a €385 million round in December 2023.
Le Chat features and pricing
The Le Chat assistant rebranded to Mistral Vibe in July 2026. This tool is a conversational assistant that provides access to Mistral models, web search, and image generation via Black Forest Labs Flux. Individual users pay $14.99 monthly for the Pro tier, while the Team tier costs $24.99 per user per month. A student plan is $5.99 per month. You should check your specific regulatory needs before choosing a provider.
| Service Tier | Monthly Price | Key Feature |
|---|---|---|
| Le Chat Free | $0 | Daily query limits |
| Le Chat Pro | $14.99 | Unlimited messages |
| Le Chat Team | $24.99 | Admin controls |
Mistral provides EU-native data processing, which helps businesses in regulated sectors like finance or healthcare meet GDPR requirements without negotiating custom enterprise agreements with US providers. The Pro tier is a paid subscription that provides unlimited messages, larger file uploads, and access to Codestral code generation. Mistral Vibe generates outputs at speeds up to 1,000 words per second. The tool supports multiple languages including French, German, Spanish, Italian, Portuguese, and Dutch. Claude requires enterprise-level contracts to achieve the same data residency levels that Mistral provides by default. The context window for Mistral Large 2 is 128K tokens, which matches Claude 3.5 Sonnet. Le Chat also is a workspace where users create and modify canvases for coding, documentation, and presentations. The platform includes a Code interpreter for scientific analysis and simulations.
Coding benchmark comparisons
No shared benchmark results exist to compare the coding accuracy of Mixtral 8x22B against Claude 3.5 Sonnet. Public evidence for coding and agentic tasks remains insufficient to name a winner between these two specific models. Mixtral 8x22B is a sparse Mixture-of-Experts model that uses 39B active parameters out of 141B total parameters and provides a 64K token context window. Claude 3.5 Sonnet is a model that maintains a 200K token context window. While Mixtral 8x22B Instruct achieves a 90.8% score on GSM8K (maj@8), Claude 3.5 Sonnet scores 79.6% on SWE-Bench Verified. Mistral Large 2 is below Claude 3.5 Sonnet on coding tasks.
Mixtral 8x22B outperforms Llama 2 70B on several reasoning and knowledge benchmarks like MMLU and HellaSwag. The model is an open-weight release under an Apache 2.0 license. Will Mistral’s open-weight strategy eventually displace the proprietary dominance of US-based labs? Mixtral 8x22B is an open-weight model.
