Baidu has announced the integration of DeepSeek into its search and AI platforms to enrich user experience. This move follows the adoption of DeepSeek models, such as DeepSeek-R1 and DeepSeek-V3, by Baidu's intelligent cloud platform, Qianfan, and the Apollo smart cabin model. Additionally, Baidu's Wen Xiaoyan has incorporated DeepSeek-R1 to enhance its photo-based problem-solving feature.
Beyond Baidu, Tencent's WeChat is also testing DeepSeek integration in its search feature, allowing selected users to access the DeepSeek-R1 model for free. Baidu continues to expand its AI model offerings, with plans to release the Wenxin 4.5 series models and make them open-source by mid-year. Baidu is also expected to launch the Wenxin 5.0 model with enhanced multimodal capabilities later this year.
Baidu's founder, Robin Li, has previously expressed skepticism about open-source models, citing the challenges in maintaining them and the potential for closed-source models to offer better commercial opportunities. He emphasized that while open-source models have academic value, the competitive market environment favors closed-source models for business success.
DeepSeek, an AI company under Quantitative Fund, gained significant attention with its January release of the DeepSeek-R1 model, which rivals the performance of OpenAI's models. The model's open-source nature and MIT license have spurred widespread industry adoption, further solidifying its influence.
