Tencent's hybrid image 3.0 model topped LMArena's global "blind test" first
2025-10-05 15:14:19
According to the official WeChat account of Tencent, the latest list of "authoritative arena" LMArena in the field of Text-to-Image shows that Hybrid Image 3.0 broke through in the world's 26 large models, topping the list, surpassing top closed-source models such as Nano Banana. According to reports, this ranking does not have any "algorithmic filters". It is selected by users around the world who compare and vote in pairs, and each vote hides the user's real experience and preferences. At present, the version of Hybrid Image 3.0 has opened the Text-to-Image capability, and versions such as graphics, image editing, and multi-round interaction will also be released in the future.
Disclaimer:
1. The information provided does not constitute investment advice. Investors should make independent decisions and bear all risks themselves.
2. The copyright of this content belongs to the original author. The views expressed herein are solely those of the author and do not represent the stance or position of this website.
Previous article：
腾讯混元图像3.0模型登顶LMArena 全球“盲测”第一Next article：
a16z：加密初创公司需采纳企业级销售策略
