According to a message released by the official WeChat account of Tencent, Tencent's self-developed Hunyuan Image 3.0 large model has successfully broken through and ranked first in the "authoritative arena" of text-to-image generation - the latest LMArena rankings. This achievement not only surpasses 26 global large models, but also leaves several top closed-source models, including Nano Banana, behind.
Tencent emphasized that this ranking does not contain any "algorithmic filter," but is entirely selected based on pairwise comparison votes from users around the world. Each vote reflects the real experience and preferences of users in actual use, making this championship title more credible.
The core breakthrough of the Huyuan Image 3.0 version lies in its high-quality image generation capability, marking a milestone victory for domestic large models in international text-to-image competitions.
Currently, the Huyuan Image 3.0 version has opened up the text-to-image generation function for user experience. According to Tencent, other advanced functions, including image-to-image, image editing, and multi-turn interaction, will be released in subsequent stages.
