Grok 2 Vision

Grok 2 vision model excels in vision-based tasks, delivering state-of-the-art performance in visual math reasoning (MathVista) and document-based question answering (DocVQA). It can process a wide variety of visual information including documents, diagrams, charts, screenshots, and photographs.