Mistral unveils ‘Pixtral 12B,’ its inaugural multimodal artificial intelligence model

The innovative French artificial intelligence startup, Mistral, has introduced its maiden multimodal creation, the Pixtral 12B, capable of managing both textual content and visual graphics, as per the report by Techcrunch. With a staggering 12 billion parameters, this model is structured upon Mistral’s Nemo 12B text-oriented model. Pixtral 12B can provide responses to image-related queries using URLs or images encoded with base64, for instance, determining the number of instances of a specific object that are discernible.

Numerous creative AI (artificial intelligence) models have undergone partial training that involved copyrighted materials, which consequently triggered legal disputes initiated by the respective copyright holders. (The AI companies, on the other hand, argue that this approach should qualify under fair use principles.)

The specific image dataset leveraged by Mistral for the development of Pixtral 12B remains shrouded in mystery.

About Author

AndyC

Andy Curtis is an award-winning security consultant, researcher and public speaker. He has been working in the computer security industry since the early 1990s, having been employed by state and federal government, leading healthcare and banking providers across three continents. He has given talks about computer security for some of the world’s largest companies, worked with law enforcement agencies on investigations into hacking groups, and is a regular voice on TV and radio explaining IT security threats.

See author's posts