Chinese AI firm SenseTime releases open source image model that generates and interprets images faster than US competitors

SenseTime released SenseNova-U1, an open source model that can generate and interpret images without translating them to text first, reducing computing power required. The model was released for free on Hugging Face and GitHub. On release day, 10 Chinese chip designers, including Cambricon and Biren Technology, announced their hardware supports U1.

SenseNova-U1 uses a technical structure called NEO-Unify that enables it to reason directly with images rather than text alone. The model is small enough to run on PCs and phones. SenseTime claims it generates higher-quality images than all other open source models currently on the market, and its main selling point is its ability to generate images much faster than competing models.

SenseTime, founded in 2014 as a world leader in computer vision, has struggled to remain competitive after ChatGPT and natural language processing systems became the industry's focus. The company began falling behind newer Chinese startups like DeepSeek and MiniMax. SenseTime made the decision last year to focus on open source, with cofounder Dahua Lin stating that "being open source or closed source is not the winning factor; the speed of iteration is."