Tencent HybriNet will release native multimodal image generation models.

Tencent's latest avatar model "Mixed Element Image 2.1" was officially released on the evening of September 9th, supporting prompt words of up to 1000 tokens and supporting native 2K avatars. In addition, Tencent also revealed that the "Mixed Element Native Multimodal Image Generation Model" is also in the works. This time, Mixed Element also open sourced a Chinese-English rewriting model - Mixed Element Text Rewriting Model, which can optimize the structure of user's text commands.