Lates News

Musk: Last week, we completed pre-training on the V7 base model, which has native multimodal capabilities. This model can directly process video/audio streams without the need to convert them to other forms for understanding. For example, it can eventually capture subtle differences in emotion and emphasis when you speak.