It's been a big — but rocky — week for AI models from China. Here's what's happened

This week has seen significant advancements in artificial intelligence from Chinese tech leaders, showcasing their growing capabilities in robotics and video generation. Major players like Alibaba, ByteDance, and Kuaishou have unveiled new AI models that highlight their ability to compete with Western counterparts, particularly in light of comments from Google DeepMind's CEO, Demis Hassabis, who noted that Chinese AI developments are merely months behind those in the U.S. Alibaba's DAMO Academy has introduced RynnBrain, a sophisticated AI model designed to enhance robots' ability to understand and interact with their physical environment. In a demonstration, a robot equipped with pincers was shown counting and picking up oranges before placing them into a basket, as well as retrieving milk from a fridge. Training these models to recognize and manipulate everyday objects is crucial, as even simple tasks can pose challenges for robotic systems. RynnBrain's unique feature includes built-in time and space awareness, allowing the robot to remember the timing and location of events, thus improving its reliability and performance in complex situations, according to Adina Yakefu from Hugging Face. On the video generation front, Seedance 2.0 has emerged as a powerful tool capable of creating realistic videos from text prompts, with the ability to incorporate existing videos and images. Creative professional Billy Boman, who has utilized Seedance 2.0 for his advertising projects, remarked on the dramatic improvements in AI video generation over the past two years, stating that the technology has evolved from producing rudimentary animations to delivering high-quality, detailed visuals almost effortlessly. Despite its successes, Seedance 2.0 has encountered challenges, such as the suspension of a feature that generated voices from user-uploaded images, following concerns about consent raised by a Chinese blogger. Kuaishou also entered the fray with its Kling 3.0 model, boasting upgrades in video consistency, photorealism, and audio capabilities, which have contributed to a notable surge in the company's stock price over the past year. Additionally, Zhipu AI has made headlines with the release of GLM-5, an open-source large-language model that claims to rival Anthropic's Claude Opus 4.5 in coding tasks, while outperforming Google's Gemini 3 Pro in certain aspects, though these assertions remain unverified. Shares of MiniMax also rose following the launch of its updated M2.5 model, which enhances AI agent functionalities designed for task automation. The rapid developments in China's AI sector signal a formidable challenge to existing Western technologies, as these firms continue to innovate and push boundaries in the field.