French AI startup, Mistral, has released its first multimodal model called Pixtral 12B. This model is capable of processing both images and text, making it suitable for tasks such as captioning images, identifying objects, and answering image-related queries. With a size of 24GB, Pixtral 12B is available for free under the Apache 2.0 license, allowing anyone to use, modify, or commercialize it without restrictions. However, web demos of the model are not yet live.
Pixtral 12B is an expansion of Mistral’s existing text-based model, Nemo 12B. By incorporating image processing capabilities, Mistral aims to enhance the functionality of its chatbot, Le Chat, and API platform, La Platforme. This integration will enable these platforms to provide more comprehensive and accurate responses to user queries.
Multimodal models like Pixtral 12B represent the next frontier for generative AI. They follow in the footsteps of tools like OpenAI’s GPT-4 and Anthropic’s Claude. These models have the potential to revolutionize various industries, from content generation to customer service. However, concerns arise regarding the data sources used to train these models.
As noted by Tech Crunch, Mistral and other AI firms likely trained Pixtral 12B using publicly available web data. This practice has sparked lawsuits from copyright holders who challenge the “fair use” argument often made by tech companies. The issue of data ownership and fair use continues to be a hot topic in the AI community, with ongoing debates about ethics and legality.
Mistral’s release of Pixtral 12B comes on the heels of the company raising $645 million in funding, which pushed its valuation to $6 billion. With Microsoft among its backers, Mistral is positioning itself as Europe’s answer to OpenAI, one of the leading players in the AI industry.
In conclusion, Mistral’s release of Pixtral 12B marks a significant advancement in AI technology. The integration of image processing capabilities into the model opens up new possibilities for applications in various industries. However, the use of publicly available web data for training raises concerns about data ownership and fair use. As the AI field continues to evolve, it is crucial to address these ethical and legal issues to ensure responsible and transparent practices.
