Command A Vision (07-2025)

Cohere's Command A Vision (July 2025) extends the Command A model family with visual understanding capabilities, enabling it to process and analyze images alongside text. It combines the strong instruction-following and enterprise capabilities of Command A with state-of-the-art vision-language modeling.

The model excels at document understanding, chart and graph analysis, image description, visual question answering, and multimodal reasoning tasks. It supports high-resolution images and can handle complex visual layouts like tables, forms, and technical diagrams.

Command A Vision is designed for enterprise applications including document processing, content moderation, data extraction from visual sources, and accessibility tools.