Alibaba (BABA) expanded its Qwen series of large language models with the launch Thursday of Qwen VLo, an upgraded artificial intelligence model that allows users to generate and edit images using text prompts and visual inputs.

Building on previous Qwen models, Qwen VLo not only understands visual content but also generates high-quality images and performs advanced edits, now available in preview via Qwen Chat for tasks like creating images from text prompts or modifying existing visuals, according to a blog post Thursday by the Qwen development team.

"This newly upgraded model not only 'understands' the world but also generates high-quality recreations based on that understanding," the Qwen team said.

Qwen VLo features a "progressive generation method" that builds images gradually for improved quality and control, while offering "precise content understanding" and support for natural language instructions to handle complex edits, style transfers, and scene reconstructions with accuracy, the team added.

The Qwen team said Qwen VLo is still in preview, with ongoing improvements planned to address current limitations in accuracy and consistency.

Price: 114.11, Change: +0.18, Percent Change: +0.16