Baidu VS DeepSeek
Chinese technology giant Baidu announced on Sunday the launch of two new artificial intelligence basic model including one that focus on reasoning capabilities.
Baidu has released a new artificial intelligence model that clarifies its reasoning methods, apparently in an effort to regroup and compete with upstarts such as DeepSeek.
The Wenxin DaMoX1 from Baidu, China’s internet search leader, works similarly to DeepSeek R1, which stunned Silicon Valley with its performance that rivals the world’s best chatbots ((19.950, -0.12, -0.60%) but is cheaper to develop. Baidu’s reasoning model excels in areas such as daily conversations, complex calculations and logical reasoning, the company said in a March 16 statement.
Baidu also upgraded its flagship base model to Wenxin Da Model 4.5. Baidu immediately made all levels of service, including the X1 model, free to its chatbot users, weeks ahead of previous plans.
The Beijing-based company pioneered chatbots modeled after OpenAI’s ChatGPT in the Chinese tech industry, but chatbots from rivals Bytedance and Moonshot AI soon became more popular. Other open-source models like Alibaba’s Tongyi Qianwen and later DeepSeek gained more recognition in the global developer community.
Baidu cited the results of multiple industry benchmark tests and said that Wenxin Grand Model 4.5 is better than OpenAI’s latest GPT 4.5 in text generation.
Baidu announced that it will open source the Wenxin model from June 30, which is another major strategic shift after the rise of DeepSeek. Baidu also integrated the R1 model into its main business, the search engine.
Baidu’s December quarterly earnings report showed a 26% jump in cloud computing revenue, reflecting a boom in generative AI. The growth, fueled by services for developers chasing computing power, was overshadowed by weak advertising sales.
In February, Baidu completed the acquisition of YY Live, a live streaming platform under Huanju Group. The $2.1 billion acquisition released about $1.6 billion that Baidu had previously deposited in an escrow account, and Baidu plans to invest the funds in artificial intelligence and cloud infrastructure. Editor/Chen Jiajing.
Chinese technology giant Baidu announced on Sunday the launch of two new artificial intelligence (AI) basic models, including one that focuses on reasoning capabilities and claims that its performance is comparable to that of Chinese AI startup DeepSeek, indicating that competition in China’s AI industry is further escalating.
DeepSeek has recently caused a stir in the international AI field with its high-performance, low-cost AI model. DeepSeek claims that the capabilities of its AI model are comparable to or even exceed those of mainstream AI models in the United States, but at a fraction of the cost of its competitors, which has had a huge impact on the AI industry landscape. Now Baidu has gone a step further, emphasizing that its new ERNIE X1 model has comparable performance to the DeepSeek R1, but costs only half the price.
The core features of ERNIE X1 include:
- Ability to use tools autonomously: The first deep-thinking AI that can use tools autonomously, emphasizing application scenarios and decision-making reasoning.
- Stronger ability to understand, plan, reflect and evolve: Baidu believes that this will make the X1 more adaptable to complex AI application needs.
Against the backdrop of increasingly fierce competition in China’s AI field, the launch of this model means that Baidu hopes to counter the rise of DeepSeek in the AI competition and consolidate its position as the leader of China’s AI. In addition to X1, Baidu also released ERNIE 4.5, an AI model with powerful multimodal capabilities. Baidu claims that it has improved language processing, logical reasoning, and memory.
Key features of ERNIE 4.5 include:
- Advanced multimodal understanding: It can process multiple types of data such as text, images, videos and audio, and can perform cross-modal content conversion.
- High EQ: Good at understanding memes and cartoons, showing a deeper understanding of language and culture.
- Improved language and creative capabilities: Better at generating high-quality text and semantic understanding than the previous generation.
As one of the first Chinese technology companies to launch ChatGPT-like AI chatbots, Baidu has been trying to compete with competitors such as OpenAI and DeepSeek in the field of AI; however, although Baidu claims that its “ERNIE” series can match the performance of GPT-4, it has failed to gain widespread market adoption, indicating that competition is still very fierce.
At present, China’s AI industry is developing rapidly and forming a competitive situation with the United States. Multimodal AI has become the focus of the next wave of competition because it can integrate different types of data and thus expand AI application scenarios.
In the future, whether Baidu can expand its market influence through these new models and fight against competitors such as OpenAI and DeepSeek remains to be verified by the market; but what is certain is that competition in China’s AI industry is becoming more fierce than ever, and this AI arms race will continue to affect the global technological development landscape.
