SenseTime, Chinese The artificial intelligence company best known for its facial recognition technology released a up-to-date open-source model on Tuesday that it says can both generate and interpret images much faster than top models developed by U.S. competitors. SenseNova U1 could assist the company regain ground after it lost its place among the top players in China’s AI development race.
The secret of the model is its ability to “read” images without having to first translate them into text, which speeds up the process and reduces the amount of computing power required. “The entire model reasoning process is no longer limited to text. It can also be based on images,” Dahua Lin, co-founder and chief scientist at SenseTime, told WIRED.
Lin, who is also a professor of information engineering at the Chinese University of Hong Kong, says models capable of direct image processing will enable future robots to better understand the physical world.
Like DeepSeek’s latest flagship model, SenseTime claims the U1 may be powered by chips made in China. “Several Chinese domestic chipmakers have completed compatibility optimization with our new model,” Lin says. On launch day, 10 Chinese chip designers, including Cambricon and Biren Technology, announced that their hardware supports U1.
This flexibility matters because U.S. export controls limit Chinese companies’ access to the world’s most advanced AI chips, especially those used for training purposes, which are currently developed mainly by Western companies such as Nvidia. “We will continue to push for training on more different chips,” Lin says. However, he also admits that SenseTime “may still need the best chips to ensure the speed of our iteration.”
SenseTime has made U1 available for free on Hugging Face and GitHub, further evidence that Chinese companies are becoming some of the most busy open-source AI developers.
SenseTime was founded in 2014 and has become a global leader in computer vision for applications such as facial recognition and autonomous driving. But as ChatGPT and other natural language processing-based AI systems became the hottest thing in the tech industry, SenseTime began struggling to make a profit and fell behind newer Chinese startups like DeepSeek and MiniMax.
SenseTime hopes that making SenseNova-U1 publicly available to anyone who can utilize it will assist it catch up with both domestic and Western AI players. Lin says the company finally decided to focus on open source last year because of the helpful feedback it receives from researchers, which allows the company to make changes more quickly. “In this day and age, being open or closed source is not a winning factor; it is the speed of iteration that counts,” explains Lin.
The move to open source software also helps SenseTime continue to collaborate with international researchers without the interference of geopolitics. The company has been repeatedly sanctioned by the U.S. government in recent years over allegations that its facial recognition technology helped power surveillance systems used to monitor and detain Uyghurs and other minority groups in China’s Xinjiang region. As a result, American companies are prohibited from investing in SenseTime and selling certain technologies to it without a license. (SenseTime has denied the allegations.)
Seeing clearly
In the accompanying technical report, SenseTime claims that SenseNova-U1 produces higher quality images than all other open source models currently on the market. Its performance is comparable to leading Chinese closed-source models such as Alibaba’s Qwen and ByteDance’s Seedream, but still lags behind industry leaders such as GPT-Image-2.0, which was released just a week ago.
However, the main advantage of this model is its ability to generate images much faster than all these models. It is based on an pioneering technical framework called NEO-Unify, which SenseTime unveiled earlier this year.