Chinese internet search giant Baidu said on Sunday that it released two new artificial intelligence models, including a reasoning-focused model, and made its AI chatbot services free as ferocious competition grips the sector.
Technology companies in China have been scrambling to release improved AI platforms since start-up DeepSeek shocked its rivals with its open-source and highly cost-efficient model in January.
Baidu announced in a WeChat post that its latest X1 reasoning model – which the company claims performs similarly to DeepSeek's but for lower cost – and a new foundation model, Ernie 4.5, were available via its AI chatbot Ernie Bot.
Baidu also made the models free to use, more than two weeks ahead of schedule. Previously, users had to pay a monthly subscription to access the company's latest AI models.
The Beijing-based company was one of China's first to roll out a generative AI platform publicly in 2023 but has struggled to gain widespread adoption, and rival chatbots from companies such as TikTok owner ByteDance and Moonshot AI have since gained more users.
Multimodal AI systems can process and integrate various types of data, including 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," Baidu said of one of the new models, as per a Reuters report.
The X1 has "stronger understanding, planning, reflection, and evolution capabilities," it said, adding that it is the first deep thinking model that uses tools autonomously.
Baidu said its latest foundation model, ERNIE 4.5, has "excellent multimodal understanding ability. It has more advanced language ability, and its understanding, generation, logic and memory abilities are comprehensively improved."
Baidu also said it has a "high EQ," and it is easy to understand network memes and satirical cartoons.
Baidu faces stiff competition in the consumer-facing AI sector, where startup DeepSeek shook up the industry at home and abroad with a model that performed comparably to competitors such as U.S.-made ChatGPT but cost much less to develop.
Since then, Chinese companies and local government agencies have rushed to incorporate DeepSeek's open-source model into their work, while other technology companies have been playing catch-up.
Baidu itself has integrated DeepSeek's R1 reasoning model into its search engine.
In February, WeChat owner Tencent released a new AI model that claimed to answer queries faster than DeepSeek, even as it incorporated its rival's technology into its messaging platform.
In the same month, Alibaba, which has partnered with Apple to develop AI for U.S.-made phones in China, announced it would invest 380 billion yuan ($52 billion) in AI over the next three years.
This month, Alibaba also released a new version of its AI assistant app, which is powered by the open-source Qwen reasoning model.
Baidu has also announced plans to follow DeepSeek's lead by making its Ernie AI models open-source starting June 30.
