(Reuters) -China’s Baidu is set to launch the next iteration of its artificial intelligence model in the second half of 2025, a person with direct knowledge of the matter told Reuters.
The new model, Ernie 5, will feature multimodal capabilities enabling it to process and convert between different formats including text, video, images and audio, CNBC reported earlier.
The planned release comes amid intensifying competition in China’s AI sector, notably from startup DeepSeek, which gained prominence after launching a reasoning model that matches OpenAI’s GPT performance at lower costs.
Despite being among China’s first movers in AI following ChatGPT’s 2022 debut, Baidu has struggled to gain widespread adoption for its Ernie large language model, even as the company claims its latest version, Ernie 4, matches OpenAI’s GPT-4 capabilities.
The tech giant’s AI offerings have lagged domestic competitors, including ByteDance’s Doubao chatbot and newcomer DeepSeek, in terms of user adoption.
Baidu CEO Robin Li told attendees in a conference held in Dubai on Tuesday that DeepSeek’s sudden emergence demonstrated the unpredictable nature of innovation.
“You just don’t know when and where innovations come from,” he said.
During the event, Li also said that investment in data centres and cloud infrastructure is still needed despite DeepSeek challenging the cost efficiency of large AI models.
(Reporting by Liam Mo in Beijing, Brenda Goh in Shanghai, Mrinmay Dey in Bengaluru; Editing by Sonia Cheema and Sam Holmes)
