By Kimberley Kao

Chinese tech giant Baidu released a new artificial intelligence model, which it claims rivals DeepSeek's AI model across various benchmarks at a lower cost.

Baidu introduced Ernie X1, a reasoning model that can support tools such as advanced search, generation of AI images, and webpage reading, in a statement Sunday. It said the model's performance is on par with DeepSeek's open-source AI model, dubbed R1, but at only half the price.

DeepSeek caused a market frenzy earlier this year with its R1 model which it claimed excels at problem solving, performing on par with OpenAI's GPT-4o model but at a fraction of the cost per use and without relying on the most advanced chips.

Baidu also unveiled Ernie 4.5, the latest version of its foundational model, that the company says outperforms OpenAI's GPT-4.5 model, based on Baidu's compiled comparison using various benchmark tests, priced at just 1% of GPT-4.5.

Also, Baidu made its Ernie AI chatbot available for free ahead of schedule.

Ernie 4.5 and Ernie X1 models will be integrated into its consumer products, including Baidu Search and its chatbot, the company said.

Originally a search-engine company, Baidu was the first in China to launch a ChatGPT equivalent, the Ernie Bot, in 2023. Baidu, along with other players in China's AI sector, face increased competition, particularly from DeepSeek.

Citi Research analysts described DeepSeek-R1 as a "disruptive debut" to the industry, and noted that companies have been announcing new releases or upgraded versions of various AI models, which may "have pushed Baidu to accelerate its release schedule,"

Using its cloud infrastructure and AI cloud's Qianfan foundation model platform, "Baidu likely will retain its competitiveness despite intensified competition among models amid lowering entry barriers," they said in a note.

Write to Kimberley Kao at kimberley.kao@wsj.com

(END) Dow Jones Newswires

March 16, 2025 23:28 ET (03:28 GMT)

Copyright (c) 2025 Dow Jones & Company, Inc.