HONG KONG — SenseTime released a new open source artificial intelligence model called SenseNova U1 on Tuesday, distributing it for free on Hugging Face and GitHub. The Chinese company, best known for its facial recognition technology, said the model can both generate and interpret images far faster than top models developed by U.S. competitors.

SenseNova U1 can read images without first converting them to text, an approach SenseTime said speeds processing and reduces computing power requirements. The company said the model can be powered by Chinese-made chips, and on release day, ten Chinese chip designers, including Cambricon and Biren Technology, announced that their hardware supports it.

"The model's entire reasoning process is no longer limited to text. It can reason with images as well," said Dahua Lin, cofounder and chief scientist at SenseTime. "Several Chinese domestic chipmakers have finished optimizing compatibility with our new model," he said.

SenseTime said it hopes that releasing SenseNova U1 publicly will help it catch up with both domestic and Western AI players. The company decided last year to focus on open source to get feedback from researchers that enables faster iteration, and said going open source helps it continue collaborating with international researchers without the interference of geopolitics. "In this day and age, being open source or closed source is not the winning factor; the speed of iteration is," he said.

U.S. export controls restrict Chinese firms from accessing the world's most advanced AI chips used for training, which are primarily developed by Western companies like Nvidia. "We will continue to push for training on more different chips," he said, adding, "We may still need to use the best chips to ensure the speed of our iteration."

SenseTime has been sanctioned repeatedly 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, U.S. firms are restricted from investing in SenseTime and selling certain technologies to it without a license. The company has denied the allegations.

In an accompanying technical report, SenseTime claimed that SenseNova U1 generates higher-quality images than all other open source models currently on the market. The company said its performance is comparable to leading Chinese closed source models like Alibaba's Qwen and ByteDance's Seedream, but still lags behind industry leaders like GPT-Image-2.0. The model relies on a technical structure called NEO-Unify that SenseTime previewed earlier this year and is small enough to run on PCs and phones.

"This is a more ambitious approach, as it still faces practical challenges," said Adina Yakefu, an AI researcher at Hugging Face. He said the technique will be especially useful in robotics, where a robot processing the visual world must sort through large amounts of information. "It has to think, 'How should I deal with all the clutter in this room? If there is a complicated machine in front of me, which button should I press?' All of these are forms of information, and they need to be integrated into the model's internal judgment," Lin said. SenseTime does not currently develop its own robots but is working with ACE Robotics, a startup led by another SenseTime cofounder.