On July 18, during the 2026 World Artificial Intelligence Conference (WAIC), SenseTime proudly introduced its premier large model, the SenseNova U1 Pro. Designed as a production-ready, native multimodal intelligent agent foundation model tailored for long-term tasks, it seamlessly blends comprehension, generation, and action capabilities. Built on a native unified architecture, the U1 Pro has undergone rigorous refinement in the realm of visual creation, boasting four pivotal features:

Firstly, its peak design aesthetics—image generation reflects professional design intuition, ensuring production-quality results. Secondly, it offers native 8K ultra-high-definition output, supporting resolutions up to native 8K, with crisp and stable text and lines even upon frame enlargement. Thirdly, it excels in ultimate graphic and detail control (exquisite graphic-text and detail management)—through enhanced integration and analysis of graphic-text information, it preserves the overall layout and content expression, with an exceptionally low error rate in text rendering. Fourthly, it supports long-term agentic closed-loop thinking—enabling sustained, interleaved graphic-text reasoning, it facilitates dozens of rounds of intelligent agent generation cycles around complex objectives, achieving synchronized and precisely controllable editing of both overall style and local text.

The U1 Pro has found applications across diverse scenarios, including office environments, e-commerce, education, and content creation, such as the automatic generation of business PPTs, corporate infographics, and product posters. Presently, the preview version is available for invitation-only testing, with the official version slated for release in August, accompanied by API services.