Baidu, the Chinese internet search giant, revealed on Sunday that it had released two new artificial intelligence models, including one focused on reasoning.
The company also made its AI chatbot services free of charge, responding to the growing competition in the AI sector.
Technology companies in China had been racing to introduce enhanced AI platforms ever since the start-up DeepSeek disrupted the industry in January with its open-source and cost-effective model.
Baidu announced via a WeChat post that its latest X1 reasoning model, which the company claimed performed similarly to DeepSeek’s model but at a reduced cost, was now available, along with a new foundation model, ERNIE 4.5.
Both models were accessible through Baidu’s AI chatbot, Ernie Bot.
The company also stated that it would offer the models free of charge, well ahead of its planned release date.
Previously, users had been required to pay a monthly subscription to access Baidu's latest AI models.
Baidu, based in Beijing, had been one of the first in China to launch a generative AI platform in 2023 but had struggled to achieve widespread adoption.
Its competitors, including ByteDance, the parent company of TikTok, and Moonshot AI, had since gained more users with their own chatbot services.
The new AI models from Baidu offer multimodal capabilities, which allow for the processing and integration of various data types, such as text, video, images, and audio, and can convert content across these formats.
“ERNIE X1 delivers performance on par with DeepSeek R1 at only half the price,” the company stated, according to a Reuters report.
Baidu explained that the X1 model featured "stronger understanding, planning, reflection, and evolution capabilities" and that it was the first deep thinking model to use tools autonomously.
Baidu also highlighted that its latest foundation model, ERNIE 4.5, demonstrated “excellent multimodal understanding ability."
It claimed that the model had more advanced language capabilities, with significant improvements in its understanding, generation, logic, and memory abilities. The company further stated that ERNIE 4.5 possessed a "high EQ" and could easily comprehend network memes and satirical cartoons.
As Baidu faced fierce competition in the consumer-facing AI sector, where DeepSeek had already shaken up the industry both at home and internationally, it was clear that the company had to adapt quickly.
DeepSeek’s model was known for delivering comparable performance to competitors like the U.S.-developed ChatGPT, yet at a far lower cost.
In response, Baidu had integrated DeepSeek's R1 reasoning model into its search engine. Meanwhile, in February, Tencent, the owner of WeChat, released a new AI model that claimed to answer queries faster than DeepSeek, although it incorporated DeepSeek’s technology into its messaging platform.
In the same month, Alibaba, which had partnered with Apple to develop AI for U.S. phones in China, announced plans to invest 380 billion yuan ($52 billion) in AI over the next three years.
This month, Alibaba launched a new version of its AI assistant app, powered by the open-source Qwen reasoning model.
Baidu also confirmed its plans to follow in DeepSeek’s footsteps by making its Ernie AI models open-source, starting on June 30.
