Baida, after

Chinese technological gigan Bake He released two new models of artificial intelligence because it was restored to return his leading position to the fiercely competitive space of the country.

Baido models launched on Sunday Included the first model focused on the company’s reasoning, as well as its first step into the open source licensing strategy.

However, experts said CNBC that while the release of models is a positive development for Baidu, they also emphasize how it plays, as it is catching up, as its Ernie Bot-one from the earliest versions in chat chatting for wide acceptance.

“The new models make Baidu more competitive, because the company lags behind the release of the model of reasoning,” said Lian Jye Su, Chief Analyst Omdia, CNBC.

The reasoning model is a big linguistic model that breaks the tasks into smaller works and views numerous approaches before generating the answer. It is designed to process complex problems in the same way to the person.

Chinese Startup DeepDeew Global AI Race DeepDeew Global and transformed China’s ecosystem in January when released its R1 reasoning, which has collected US competitors despite the share.

Baida is – Note Its new ERNIE X1 reflection model “provides performance along with Deepseek R1 only half the cost” and has “opportunities for a stronger understanding, planning, reflection and evolution.” CNBC failed to check this statement on its own.

According to Wei Song, the chief analyst of artificial intelligence in the Controppoint Research, the future Baid competitiveness may depend on whether its new models provide for the promised results and costs.

“Baidu is clearly in catch-up mode, largely due to the slow pace of innovation and underestimation of fast shifts in the dynamics of the market,” Song said.

However, despite The initial impulseSince then, product Baidu Ernie dimmed with competitors including startups as well as large technology companies such as Alibaba and byte.

Experts list a number of reasons for fighting Baida and slowly the speed of innovation.

“Baidu lagging behind when they tried to build their own models and compete for AI financing,” said Ray van, the chief analyst and founder of the Constreelation Research, CNBC. He added that the company also suffered from recent government repression and distracted by “normative nonsense”.

Own models support their source code and basic architecture confidential, unlike models similar to Deepseek, the source code is provided in an open network for possible modifications and redistribution.

“Using a closed source approach means that (Baidu) trained its model from scratch, whereas the open source models were able to use some parts that were utilities,” said Kai van, senior analyst Morningstar.

The latest AI Baidu models, however, are a shift in the company’s strategy to open and free models.

“Baidu has always supported his own business -model and was voice against open source, but violators such as Deepseek have proven that open source models could be as competitive,” Su said.

He added that Baida “just goes the step” of his largest competitors in China, namely Alibaba, Deepseek and Tencent, which all now released open source models.

Tape of Ann Profit calls late last yearBaidu CEO Robin Lee said he expects the Generative II to turn the Baid into a “new app -II app”.

“The secret of AI is data, chips, good math and cheap energy … You have data to play. They have data,” said Wang Constreelation Research.