The race to create powerful and low-cost artificial intelligence models is heating up. Weeks after DeepSeek rattled the AI landscape with its affordable R1 model, Baidu (BIDU) has also followed suit with a new AI model that it claims is better across various benchmarks in addition to being affordable.
Baidu AI Model
Ernie X1 is the new AI model sending shockwaves owing to its ability to support tools across advanced search and enable the generation of AI models and webpage reading. Along with releasing Ernie 4.5, the most recent iteration of its foundational model, which the company claims outperforms OpenAI’s GPT-4.5 model based on comparing the two models using a number of benchmark tests. Baidu also made its Ernie AI chatbot available for free ahead of schedule. The Ernie 4.5 and Ernie X1 models will be incorporated into Baidu’s consumer products, such as Baidu Search and its chatbot.
While the Erne is at par with DeepSeek’s R1 model on performance, Baidu insists it was developed at half the price. The remarks come when tech giants are trying to outdo each other in the development of AI models that can perform various functions at the lowest cost possible.
China vs. US
The development has rattled the semiconductor sector with chip giants coming under pressure having become clear that it’s possible to develop AI models without spending much on chips. Nvidia is one of the companies that have come under scrutiny in the aftermath of DeepSeek revelations. There are growing concerns that the sale of the company’s powerful and expensive Blackwell chips could come under pressure as the focus turns to developing cost-effective AI models.
China has the upper hand in developing powerful AI models at almost half the cost. DeepSeek Baidu and Alphabet (GOOGL) have already set the ball rolling with the launch of powerful AI models rivaling the likes of OpenAI’s ChatGPT. Similarly known Chinese startup Manus AI launched a preview version of an AI agent capable of screening resumes, creating trip itineraries and analyzing stocks in response to basic instructions from the user.
The milestones achieved by Chinese companies in developing AI models should continue to trigger questions about the US lead in developing AI models and innovations. That’s because US companies have always seen this as a key investment area.
Google AI Robot
Meanwhile, Alphabet’s Google is integrating its DeepMind artificial intelligence technology models into physical robots. The company has unveiled two new AI models, Gemini Robotics and Gemini Robotics-ER. These models operate on Gemini 2.0, which is Google’s most potent AI to date. Gemini Robotics goes beyond outputs like text and images, where generative AI has flourished up to this point, and into physical action commands to control robots.Google will collaborate with robotics developer Apptronik, which has previously worked with Nvidia (NVDA) and NASA, to create the next generation of humanoid robots using Gemini 2.0. In demonstration videos, Google showed Apptronik robots, outfitted with the new AI models, plugging something into a power strip, packing a lunchbox, moving plastic vegetables, zipping up a bag, and zipping up a lunchbox.
