• The Alibaba International AI team introduced Ovis-U1, a powerful multimodal AI model capable of advanced image generation, editing, and cross-modal tasks.
  • Ovis-U1 is open-source, enabling global developers to build upon it. This aligns with Alibaba’s $53 billion AI and cloud investment plan for the next three years.

The new multimodal big model Ovis-U1 was formally introduced by the Alibaba International AI team on June 29, 2025, signalling yet another advancement in multimodal artificial intelligence (AI).

Ovis-U1, the most recent masterpiece in the Ovis series, combines multimodal comprehension, image production, and editing skills.

It exhibits strong cross-modal processing capabilities and opens up new avenues for industry applications, researchers, and developers.

Ovis-U1 is a useful tool for educators, designers, and advertisers because it outperforms standard AI models that only handle one kind of data.

It is distinct from competitors in that it can execute complex visual editing instructions within one command without losing semantic precision.

Due to Alibaba's commitment to open-source technology, Ovis-U1 can be accessed and built upon by developers globally, promoting collaborative innovation among the community of AI professionals.

The announcement complements Alibaba's ambitious AI strategy, backed by a $53 billion investment in cloud infrastructure and AI over the next three years.

Alibaba's Qwen series has been a success, and Ovis-U1 inherits the multimodal strengths of Qwen-VLo and its successors.

Ovis-U1's potential to transform AI development is evident in X posts where users commend its integrated strategy for multimodal tasks.

Alibaba has become the leader of the global AI community through Ovis-U1, and Chinese technology giants such as ByteDance and SenseTime are competing to develop similar models.

Alibaba is committed to ongoing enhancements, even though the model remains in preview and might sometimes be flawed.

Ovis-U1 is now available on platforms such as Hugging Face, which shows that a new era of high-performance, accessible AI solutions is on the horizon.

Edited by Annette George