AI Photo: VCG

 
Ant Group on Tuesday launched a next-generation multimodal artificial intelligence (AI) assistant, the first AI assistant capable of generating multimodal content with full-code implementation. The launch marks another major AI push from Alibaba's ecosystem following the Qwen app, according to a statement the company sent to the Global Times.

The new app, named Lingguang, can generate fully coded mini-apps on mobile devices in as little as 30 seconds through natural-language prompts, producing outputs that are editable, interactive and shareable. It is the industry's first AI assistant capable of full-code multimodal generation, launching with three core features — Lingguang Chat, Lingguang Flash Apps and Lingguang Vision — supporting 3D, audio-video, charts, animation, maps and other multimodal formats, the company said.

Recent launches from Ant Group, Baidu and iFlytek, each rolling out new AI products, show that China's tech giants are converging in the same direction of innovation. They are also intensifying efforts to capture domestic user traffic in a market that is increasingly contested by leading overseas AI platforms, experts told the Global Times.

Designed around three key features, Lingguang aims to make complex knowledge simpler and more intuitive, transforming how users access information and interact with AI tools.

For example, in an education setting, when a user asks Lingguang a knowledge-based question, the assistant can identify and distil the key concepts, present them in a clear, structured way, and generate 3D animated models or interactive tables, making complex information easy to grasp at a glance.

A key innovation is that Lingguang can build and instantly run custom "flash apps" directly within a conversation in as little as 30 seconds. The feature makes it the industry's first general-purpose AI tool that allows non-technical users to create functional applications with ease, the company said.

Also, Lingguang provides real-time understanding of complex scenes and dynamic footage captured by users, delivering contextual insights and enabling on-the-fly editing or generation of images and videos through commands, the statement said.

The Lingguang release came just a day after Alibaba unveiled its Qwen initiative as it pushes aggressively into the AI-to-consumer market. A trial version of the Qwen app was launched on the same day. Powered by Qwen3, the app aims to compete directly with OpenAI's ChatGPT by offering free access and embedding itself into various daily-use ecosystems.

"Chinese tech firms rolling out new AI products at the same time shows a shared judgment about where the global industry is heading," Liu Dingding, a veteran industry analyst, told the Global Times on Tuesday. "AI competition is inherently global. Chinese companies are benchmarking against top US models, and the gap is narrowing quickly. Years of overseas experience by Tencent, ByteDance and Alibaba gives them strong advantages in localization and user insight."

"With underlying models now much stronger, companies are putting far more emphasis on user experience and real, workable applications," Liu said.
Liu said that Chinese firms are well-positioned to compete globally. "Major Chinese platforms have solid technical foundations and large global user bases. As these strengths come together, Chinese AI companies have a growing opportunity to stand alongside US giants on the world stage."

A recent industry report showed that as of June, China had 515 million generative AI users, up 266 million from December 2024, the People's Daily's reported.

The report showed that China has released more than 1,500 industry-specific models so far, covering 50 key sectors and more than 700 application scenarios. Generative AI products in China are becoming increasingly mature, with more than 90 percent of users choosing domestic large models as their first option, according to the People's Daily.