Mistral AI has officially launched Mistral Large 4, also known as Le Chonk, as a public preview. This groundbreaking model features an impressive 1.05 trillion parameters and utilizes a Mixture of Experts (MoE) architecture, with 49 billion parameters active at any given time. The model is capable of processing native image inputs and supports a context window of up to one million tokens.
Trained on 3,800 NVIDIA Grace Blackwell GPUs housed in Mistral’s European data centers, Mistral Large 4 is now accessible via API, with open weights expected to be released soon. This model represents a significant leap in AI capabilities, particularly in multimodal tasks, and is anticipated to set new benchmarks in the field.
Compiled automatically by the Tech AI Newsdesk from public AI-news sources and summarised in our own words.
