TMTPOST -- Baidu's CEO has acknowledged that the rise of DeepSeek played a key role in the company's decision to open-source its own AI model. Open-source models allow for free and unrestricted access to software, enabling users to modify and distribute it for various purposes. DeepSeek, a rapidly growing Chinese startup, made waves in January when it launched an open-source AI model that quickly gained attention for outperforming leading models from OpenAI, Meta, and other top tech firms, all while reportedly being developed at a lower cost. "One thing we learned from DeepSeek is that open-sourcing the best models can significantly accelerate adoption," Robin Li, Baidu's CEO, said during an earnings call on Tuesday. "When a model is open-sourced, people are naturally curious to try it, which helps drive wider usage." Baidu announced last week that it will open-source its latest AI model, Ernie 4.5, starting June 30, marking a significant shift from Li's previous stance that large language models should not be open-source. In the past, Li had argued that closed-source models tend to be more powerful and cost-effective. In a 2024 interview with Chinese outlet Yicai, he stated that open-source models were better suited for academic purposes rather than commercial use. Other industry leaders, including the CEO of Mistral and Meta's chief AI scientist, have also hailed DeepSeek's success as a major win for the open-source movement.