On Sunday, Baidu (9888.HK) announced the introduction of two new artificial intelligence models, one of which is designed for advanced reasoning and is said to compete with DeepSeek’s offerings. This move comes as Baidu seeks to distinguish itself in the highly competitive AI landscape.
DeepSeek, a Chinese AI startup, has recently launched its own models, claiming they match or even surpass the leading AI technologies from the United States, all at a significantly lower cost. This development has shaken the industry and revitalized the global AI competition.
Baidu highlighted that its ERNIE X1 model performs similarly to DeepSeek’s R1 but is available at just half the price. The X1 boasts enhanced capabilities in understanding, planning, reflection, and evolution, and is noted as the first deep thinking model capable of using tools independently.
Additionally, Baidu introduced its latest foundation model, ERNIE 4.5, which features exceptional multimodal understanding. This model has improved language skills and enhanced abilities in understanding, generating, logic, and memory.
Baidu also emphasized that ERNIE 4.5 possesses a high emotional quotient (EQ), making it adept at interpreting internet memes and satirical cartoons.
As one of the first major tech companies in China to launch a ChatGPT-style chatbot, Baidu has faced challenges in achieving widespread use of its Ernie large language model, despite asserting that its performance is on par with OpenAI’s GPT-4, amid intense competition.
Multimodal AI systems are designed to handle and integrate various data types, including text, video, images, and audio, allowing for seamless content conversion across these formats.
