Mistral Large is the French AI lab's top-tier proprietary model, built for general reasoning, coding, and multilingual tasks, and positioned as Mistral's answer to GPT-4-class and Claude-class models from the larger US labs. It's available through Mistral's own API and through major cloud providers, giving European companies in particular a frontier-capable model option from a lab headquartered in the EU, which matters for some data residency and regulatory considerations.
Mistral built its early reputation on Mixtral, an open-weight mixture-of-experts model family that was genuinely competitive with much larger dense models while being cheaper to run, and that open-weight lineage still shapes how the company is perceived even as its most capable models, including Mistral Large, have moved to a more traditional closed, API-first release model.
Pricing is usage-based through its API, with a free tier for experimentation, and Mistral continues to release some models, like Codestral, under genuinely open licenses even as its flagship stays closed, a dual strategy that keeps it relevant in both the open-model and frontier-model conversations at once.