China’s Baidu launches two new AI models as industry competition heats up

On Sunday, Baidu, China’s internet search provider, announces launching two other artificial intelligence models, one being a new reasoning-focused model, claiming it equal to the model presented by DeepSeek; it promises to show off some skills at a time of high competition in AI.
The launch of AI models by the Chinese startup DeepSeek that it claims are as good as, or perhaps even better than, the best-performing models in the United States, while also being available at a fraction of the cost, has shaken the industry and renewed the global AI race.
One of its new models, ERNIE X1, is claimed by Baidu to perform just like DeepSeek R1, at half the price. It is endowed with “greater understanding, planning, reflection, and evolution capabilities”; it is also the first deep thinking model that autonomously applies tools.

ERNIE 4.5, according to Baidu, is endowed with “superior multimodal understanding ability. It enjoys an advanced language ability, with improvements in understanding generation, logic, and memory specifically.

” It also has “high EQ,” according to Baidu, and it can grasp network memes and satirical cartoons easily. One of the earliest tech giants in China to release a ChatGPT-like chatbot, Baidu has faced a long struggle to gain traction for its Ernie large-language model, even while claiming that it performs on par with OpenAI’s GPT-4 amidst cutthroat competition. Multimodal AI systems process and fuse different data forms-text, video, images, and audio-and they can coerce content through the formats.