xAI announced the launch of its new multimodal model, Grok-1.5V (Grok-1.5 Vision), which represents a major step forward in the development of truly intelligent systems. The company says Grok-1.5V goes beyond text understanding to integrate vision capabilities, allowing it to understand documents, diagrams, charts, photos, and more.
According to xAI, Grok-1.5V outperforms leading competitors on key benchmarks. It excelled on xAI"s RealWorldQA dataset, an evaluation of over 700 real-world images paired with questions. This dataset evaluates an AI"s ability to understand complex visual scenarios and measures progress toward general intelligence.
We are particularly excited about Grok’s capabilities in understanding our physical world. Grok outperforms its peers in our new RealWorldQA benchmark that measures real-world spatial understanding.
Grok-1.5V"s multifaceted perception was demonstrated through examples such as transforming a child"s drawing into a bedtime story. It also explained memes, converted tables to CSV format, and diagnosed problems with wood decking based on images alone. xAI believes that such diverse tasks demonstrate Grok-1.5V"s potential for a wide range of applications.
In case you missed it, Elon Musk launched xAI in July 2023. Musk said the goal of the new company is to create more transparent and accountable AI that can work better with humans.
One of the big issues surrounding AI right now is the seriousness with which everyone is taking it. Some say it will be the end of jobs, others hate its artistic abilities and claim it"s not really art, and others complain that schoolchildren shouldn"t use it for homework.
In the coming months, the company expects to focus its research on several key areas. xAI aims to improve the visualization and explanation of Grok"s multimodal reasoning.
xAI will soon be looking for the first testers of Grok-1.5V to provide feedback and help improve its multimodal reasoning. Access to the Grok beta is initially limited to X"s Premium+ subscribers, who receive additional benefits and support.