Tencent Holdings has launched Hy Image 3.5 Preview, its latest artificial intelligence model for image generation, as the Chinese technology company expands its generative AI capabilities amid growing competition from rivals including ByteDance and Alibaba.
Released on September 22, the new model supports text-to-image and image-to-image generation, along with multi-turn conversational editing. This allows users to generate an image and continue refining it through subsequent instructions instead of starting the creation process again.
Tencent said Hy Image 3.5 Preview improves on its previous-generation image model, with upgrades across areas such as text rendering, visual layouts, photorealism, stylistic expression and consistency during image editing. The company positions the model as a professional-grade visual generation system aimed at supporting more practical content creation workflows.
The model can also work with reference images, allowing users to provide visual inputs alongside text instructions. This broadens its potential applications across marketing, design, advertising and other creative workflows where maintaining visual consistency across multiple outputs can be important.
Tencent is integrating Hy Image 3.5 Preview across several of its products. These include its Yuanbao AI assistant as well as creative, productivity and design tools within the company's broader ecosystem. The model is also being made available through Tencent Cloud, extending access to businesses and developers building AI-powered applications.
The launch is part of Tencent's wider effort to expand its Hy, also known as Hunyuan in China, family of artificial intelligence models beyond text-based interactions. The company has been developing models spanning reasoning, agents, image generation, video and other multimodal capabilities.
Hy Image 3.5 Preview arrives as competition in China's generative AI market continues to increase. Technology companies including Alibaba and ByteDance have been investing in models and applications spanning image and video generation, enterprise AI and consumer-facing assistants. Alibaba is also expanding its AI infrastructure and Qwen model portfolio, while ByteDance has built a growing suite of generative AI products.
For marketers and creative teams, improvements in image-generation models could expand the use of AI for producing campaign concepts, product visuals and other digital assets. However, the practical adoption of such tools will also depend on factors including output quality, brand consistency, cost and integration with existing creative workflows.
Tencent's latest release signals a further expansion of its AI portfolio as China's major technology companies compete to bring increasingly capable generative models into consumer and enterprise products.
Disclaimer: This article may include information derived from interviews, press releases, public statements, research, company communications and other publicly available or third-party sources. Such material may be summarised, paraphrased or contextualised for journalistic and editorial purposes. All rights in third-party content remain with their respective owners.