Baidu, a leading player in China's tech landscape, is gearing up to launch the next generation of its artificial intelligence model, Ernie 5.0, in the latter half of this year. Although the company has not made an official announcement, insiders reveal that Ernie 5.0 will feature "big enhancements in multimodal capabilities." This move highlights Baidu's ongoing commitment to advancing its AI technologies amidst stiff competition from both domestic and international players.
The Ernie series, Baidu's flagship AI model, functions as a "foundation model" capable of understanding language, generating text and images, and facilitating natural language communication. Since the debut of its ChatGPT-like chatbot, Ernie, in March 2023, Baidu has been at the forefront of integrating generative AI into its consumer and business products. Despite initial success, the Ernie product has faced challenges, being overshadowed by other AI chatbots emerging from major Chinese tech firms and startups.
Currently, the Ernie model is in its fourth iteration. The original Ernie 4 was launched in October 2023, followed by an enhanced "turbo" version in August 2024. Baidu's anticipated release of Ernie 5.0 comes at a crucial time as Chinese companies race to develop cutting-edge AI models to keep pace with U.S.-based competitors like OpenAI. The latest version of OpenAI's ChatGPT, GPT-4o, was introduced in May 2024, but no timeline has been set for GPT-5's release, as confirmed by OpenAI CEO Sam Altman.
Baidu's AI advancements extend beyond chatbots. The company's Wenku platform, designed for creating presentations and documents, has seen significant growth. By the end of 2024, Wenku amassed 40 million paying users, marking a 60% increase from the previous year. Enhanced features like AI-generated presentations from financial filings were rolled out to users starting January, showcasing Baidu's drive to integrate AI across various applications.
Baidu CEO Robin Li has been vocal about the transformative potential of AI innovation. Speaking at the World Governments Summit in Dubai, he emphasized the economic impact of reducing inference costs for foundation models.
"We are living in an exciting time… The inference cost [of foundation models] basically can be reduced by more than 90% over 12 months," – Robin Li
This reduction translates to increased productivity, underscoring the essence of innovation.
"If you can reduce the cost by a certain percentage, then that means your productivity increases by that kind of percentage. I think that's pretty much the nature of innovation," – Robin Li
Baidu's strategic focus on enhancing multimodal capabilities with Ernie 5.0 reflects its ambition to reclaim leadership in the AI sector. As global competition intensifies, Baidu's efforts are indicative of a broader trend among Chinese tech companies striving to match and surpass their Western counterparts.