SenseTime has introduced the production version of its image creation model, SenseNova U1 Pro. This model is now available to users through the company's Raccoon application and also via the SenseNova API service. According to a report by TechNode on September 22, 2026, the information was sourced from an announcement published on Sina Finance. The official English name for the product is SenseTime / SenseNova.
This release differs from previous coverage of the SenseNova U1.5 research and open-source training code. SenseTime asserts that U1 Pro utilizes a reasoning chain alternating with a text-to-image process to create visual materials containing structured information and readable text directly on the canvas, as well as production-ready layouts.
Instead of treating captions and pixels as separate stages, the model integrates reasoning steps with image formation. This ensures that the layout hierarchy, typography placement, and graphic elements remain consistent throughout the generation process. The product supports outputting images with resolutions up to 8K and allows for custom aspect ratios, targeting large formats and high detail rather than just square crops for social media.
The announcement lists target use cases, including infographics, posters, architectural visualizations, and other content that requires a uniform style and composition across the entire finished piece. These scenarios highlight the necessity of controlling composition and text rendering within the image itself, rather than using a separate layout tool after generation is complete. Thus, marketing and design teams can request production-ready frames from a single prompt.
Distribution occurs through two channels: consumers and creators gain access via Raccoon, while product teams can utilize programmatic access through the SenseNova API to integrate generation into their workflows. SenseTime has not published a full public model card detailing parameters, licensing terms, or independent benchmarking tables in the TechNode summary in English. Therefore, buyers are advised to view the claims regarding the 8K limit and layout capabilities as company assertions until third-party evaluations are available. For buyers of multimodal systems tracking Chinese image models, the nearest signal is the availability of the ready-made U1 Pro SKU for production in the Raccoon and API channels with an explicit ultra-high resolution limitation.
