Tencent Releases Hunyuan Image 3.5 Preview, Overall Capabilities Improved by Approximately 30% Compared to Previous Generation

According to TechFlow, on September 22, Tencent officially released the professional-grade image generation model Hunyuan Image 3.5 Preview (Hy Image3.5 preview). The model supports text-to-image and image-to-image generation, allows uploading up to 5 reference images per request, supports various aspect ratios, outputs up to 2K resolution, and enables multi-turn conversational creation.

Tencent stated that the model has demonstrated significant improvements in text rendering and layout, photorealism and stylistic expressiveness, and editing consistency, with comprehensive capabilities increased by approximately 30% compared to Hunyuan Image 3.0. Currently, the model has been integrated into products such as Yuanbao, WorkRally, OnSolo, Miora, WorkBuddy, and ima, and is simultaneously available through Tencent Cloud services; the Tencent Cloud API charges 0.15 RMB per 2K image generated, with billing applied exclusively to the output images.