xAI releases Grok-1.5V, a multimodal AI model with vision support

xAI, an artificial intelligence company under Musk, announced the launch of its first multimodal AI model, Grok-1.5V. In addition to its powerful text processing capabilities, Grok can also process various visual information, including documents, charts, screenshots, and photos. In benchmark tests in multiple fields, Grok-1.5V's performance is comparable to existing cutting-edge multimodal models. Especially in xAI's newly launched RealWorldQA benchmark test, Grok surpassed similar models in its ability to understand the real-world space. The RealWorldQA dataset contains more than 700 images and aims to evaluate the basic understanding ability of multimodal models in the physical world. Grok-1.5 will soon be open to early testers and existing users.