Mistral AI drops new ‘mixture of experts’ model with a torrent link

As Google unleashed a barrage of artificial intelligence announcements at its Cloud Next conference, Mistral AI decided to jump into action with the launch of its latest sparse mixture of experts (SMoE) model: Mixtral 8x22B. However, unlike its competitors, the Paris-based startup, which raised Europe’s largest-ever seed round in June 2023 and has become a rising star in the AI domain, didn’t push the release through a demo video or blog post. Instead, it has continued with the unorthodox approach of dropping a torrent link on X to download and test the new model. This marks the third major model release in the last couple of days, following the general availability of GPT-4 Turbo with vision and Gemini 1.5 Pro. Meta has also dropped hints of releasing Llama 3 next month. While the capabilities of Mixtral 8x22B remain undisclosed, AI enthusiasts were quick to point out that this is a big release from Mistral – with the torrent leading to four files totaling 262GB – and might be difficult to run locally. “When I bought my M1 Max Macbook I thought 32 GB would be overkill for what I do since I don’t work in art or design. I never thought my interest in AI would suddenly make that far from enough,” a Reddit user wrote when discussing the new model. Soon after Mistral’s X post to download Mixtral 8x22B, the model also dropped on Hugging Face, where users can use it for further training and deployment. The listing notes that since the model is pretrained it does not have any moderation mechanisms too. Together AI has also made the model available for users to try. With its sparse MoE approach, Mistral strives to provide users with a combination of different models, each specializing in a different category of tasks, to optimize for performance and cost.

Full story : Paris-based Mistral AI launches Mixtral 8x22B, its latest sparse mixture of experts model, after releasing Mixtral 8x7B in December 2023.