Baidu has announced the introduction of two new multimodal AI models, Ernie 4.5 Turbo and Ernie X1 Turbo, engineered to process a variety of media types including text, audio, images, and video. These innovative models are designed to surpass the limitations of traditional AI by offering enhanced accuracy and cost-efficiency. The models are part of Baidu's efforts to boost its artificial intelligence capabilities amidst rising competition.
In tandem with the launch of these AI models, Baidu has deployed a robust AI cluster consisting of 30,000 P800 Kunlun chips. According to Reuters, this infrastructure significantly empowers Baidu to train AI models with hundreds of billions of parameters and allows customized model tuning for numerous clients concurrently. This powerful setup is critical for supporting the company's pursuit of advanced AI applications.
Baidu is also making a strategic pivot towards open-source AI development, promising to make its Ernie 4.5 model open source by June 2025. This move marks a notable departure from Baidu's previous closed-source strategy and is expected to enhance collaboration and innovation within the AI community. The company's focus on practical AI utilization, alongside this shift, aims to solidify its competitive footing against both domestic and international tech giants.