Baidu, China's tech giant, is set to unveil the next version of its AI model, Ernie 5, in the second half of 2025. This new iteration will introduce multimodal capabilities, allowing it to process and convert a variety of formats such as text, video, images, and audio, offering significant advancements in AI technology.

The release is timed to respond to increasing competition in China's rapidly developing AI sector, particularly from the startup DeepSeek. The company has gained attention with a reasoning model that rivals OpenAI's GPT while offering lower costs. Despite being one of the first to enter the AI space in China after the debut of ChatGPT in 2022, Baidu has faced challenges in gaining widespread adoption of its Ernie model.

Baidu's AI offerings have struggled to keep pace with competitors like ByteDance's Doubao chatbot and DeepSeek in terms of user uptake. The company maintains that its latest version, Ernie 4, is comparable to OpenAI's GPT-4, but the adoption rate has been slower than anticipated. CEO Robin Li acknowledged the unpredictable nature of innovation, noting DeepSeek's rise as a reminder that the future of AI is uncertain and can emerge from unexpected places.

Li also stressed the importance of continued investment in data centres and cloud infrastructure to keep up with the evolving demands of AI, even as newer models, such as DeepSeek's, challenge the cost-efficiency of large AI models.