On January 22, 2026, Baidu officially launched its indigenous full - modality large - scale model, Wenxin 5.0. This remarkable model comes with a staggering 2.4 trillion parameters. It leverages cutting - edge indigenous full - modality unified modeling technology, enabling seamless support for a variety of information input and output formats, including text, images, audio, and video.

For individual users, the Wenxin APP and the official website of Wenxin Yiyan offer a gateway to experience this advanced model. Meanwhile, enterprises and developers can tap into its capabilities through Baidu's Qianfan platform. In more than 40 authoritative benchmark evaluations, Wenxin 5.0 has demonstrated outstanding performance. Its language and multimodal understanding abilities have outshone those of well - known models such as Gemini - 2.5 - Pro and GPT - 5 - High, firmly establishing it as a top - tier contender on the international stage.