HONG KONG, March 16 (Reuters) – Chinese tech giant Baidu announced on Sunday the launch of two new artificial intelligence (AI) models, including one designed for advanced reasoning, which it claims rivals DeepSeek’s latest model. This move highlights Baidu’s efforts to stand out in an increasingly competitive AI landscape.
DeepSeek, a rising AI startup in China, has disrupted the industry by introducing AI models that it claims can match or even outperform leading U.S. models at a fraction of the cost. This has reignited the global AI race, prompting established players to accelerate their innovations.
"ERNIE X1 delivers performance comparable to DeepSeek R1 at just half the cost," Baidu stated. The X1 model is said to have enhanced capabilities in comprehension, planning, reflection, and autonomous development. It is also described as the first deep-thinking AI model capable of independently utilizing tools.
Baidu’s second new model, ERNIE 4.5, is built for superior multimodal understanding. The company claims it features significant improvements in language comprehension, content generation, logical reasoning, and memory retention.
Additionally, ERNIE 4.5 is said to possess high emotional intelligence, allowing it to interpret internet memes and satirical cartoons with greater accuracy.
Despite being one of the first Chinese firms to launch a ChatGPT-style chatbot, Baidu has struggled to achieve widespread adoption of its Ernie large language model. While the company asserts that Ernie’s capabilities rival OpenAI’s GPT-4, intense competition in the AI industry remains a formidable challenge.
Multimodal AI systems like ERNIE 4.5 can process and integrate various forms of data—including text, images, video, and audio—while seamlessly converting content between these formats. With these latest advancements, Baidu aims to strengthen its position in the global AI race.