According to Jiemian News, CITIC Construction Investment's research report on October 9 pointed out that the global AI field is progressing at a rapid pace, and various manufacturers continue to make efforts: Microsoft released GPT-4V and released a technical report, focusing on the effectiveness of GPT-4V in performing various tasks. , and ultimately believed that GPT-4V has unprecedented capabilities in processing arbitrary multi-modal inputs and has huge application potential in a variety of scenarios; Meta launched a multi-modal large model AnyMAL based on LLaMAv2 to achieve the unification of images, audio, text and other modalities , while raising the multi-modal model effect to a new level; Google released one of the world's largest general-purpose large models "RT-X" and opened the training data set Open X-Embodiment. The field of robot models may usher in an ImageNet moment. We recommend continuing to pay attention to the development and investment opportunities of AI large model technology, especially the opportunities brought by multi-modal models and robot models.