Mistral AI has unveiled a new free tier to give developers the opportunity to experiment with and fine-tune their AI models, as announced in a recent blog post. This new offering allows developers to create and test applications using the startup’s AI models at no cost, marking a significant step in Mistral’s efforts to attract developers. In addition, the company has reduced prices for accessing its AI models through API endpoints and introduced image processing capabilities to its free consumer AI chatbot, le Chat.
The Paris-based startup, now valued at $6 billion, is following a growing trend in the AI industry: providing more features for less. As advanced large language models (LLMs) become more widespread and accessible in the developer community, Mistral aims to remain competitive against other big players like OpenAI, Google, and Anthropic, all of which also offer free tiers with limited usage for developers.
Mistral’s free tier is available through its API-serving platform, known as “la Plateforme,” and is designed to help developers explore, evaluate, and prototype AI models without any initial costs. However, developers looking to move beyond experimentation and use Mistral’s models in production will need to upgrade to a paid commercial tier, which offers higher rate limits for more intensive use.
While Mistral positions its AI models as “open,” the company acknowledges that hosting and managing the necessary infrastructure to run these models can be complex for developers. By offering a free tier through la Plateforme, Mistral simplifies the process, making it easier for developers to start building with their tools. This approach also creates an opportunity for the company to upsell premium features, such as higher rate limits, to developers as they scale their projects.
In addition to the free tier, Mistral has significantly reduced the prices of its API endpoints. Access to Mistral NeMo, Mistral Small, and Codestral models is now more than 50% cheaper, while Mistral Large has seen a 33% price reduction.
The company has also rolled out its first multimodal AI model, Pixtral 12B, which processes both images and text. Now available on le Chat, this model enables users to scan, analyze, and search image files. Pixtral 12B was introduced last week as part of Mistral’s ongoing efforts to expand its AI capabilities.
