META The Openai key

Meta has rented the very influential Openai researcher, Trapit Bansal, to work on male thinking models under The new SuperInteigence of the company’s new SuperINGENGE AI of the companyA person familiar with it tells Techcrunch.

Openai spokeswoman Kayla Wood has confirmed that Bansal has left Openai. Banusal LinkedIn page He says he left Openai in June.

Bansal has worked in Openai since 2022 and has been a major player in starting the company’s work on reinforcement learning alongside co -founder Elijah Sutsv. It was listed as a basic contribution to the first thinking model in Openai, O1.

Bansal can offer a large elevator to the SuperIntelligence Laborat AI AILEXANDR WANG Executive Director and Looking to add Former CEO of GitHub Nat Friedman and co -founder of Safe Superinteldance Daniel Gross. Bansal Meta can develop the AI ​​Frontier thinking model competing with the leading technology in industry, such as Openai’s O3 or Deepseek’s R1. Currently, Meta does not offer a model for the eminent organization.

In recent months, Mark Zuckerberg has been in an employment to create a new artificial intelligence team in Meta, Providing compensation packages of $ 100 million For senior researchers who join his company. It is not clear what Bansal offered to join this deal.

However, it seems that Zuckerberg has succeeded in getting the best research talent from artificial intelligence.

Three other former researchers at Openai – Lucas Bayer, Alexander Colinnikov, and Chiwaha Chai joined – Wall Street Magazine It was mentioned on Wednesday. Bansal will join them, along with the former Jack Rae Rae Researcher and former Sesame Startup, Johan Schalakwyk, according to Bloomberg.

For more filling the new artificial intelligence unit, Zuckerberg tried to Obtaining startups with heavy artificial intelligence laboratoriesLike Sutskever’s Safe Superinment, Mira Murati’s Thinking machinery laboratories and confusion. However, these talks did not advance to the final stage.

In Bodcast recently, the CEO of Openai Sam Altman said that Meta is trying to hunt the talent of his higher youth, but he claimed that.Not one of our best people They decided to take it on that. ”

A Meta spokesman refused to comment.

Thinking models of artificial intelligence offers a major field for the Meta’s Ai Superintecess Correction. Last year, Openai, Google and Deepseek shipped very logical thinking models that have pushed the limits of what the program could do. By training artificial intelligence models to work through problems before giving an answer, using additional time and resource calculation to do so, artificial intelligence laboratories have found success in improving the performance of artificial intelligence in the real standards and tasks.

The Meta’s Ai SuperIntelligence can become a major internal collection that operates products throughout the company, such as Google’s Deepmind unit. Meta has an ambitious effort for Building artificial intelligence agents to work Under the former CEO of sales in artificial intelligence, Clara Shih. In order to build competitive factors, Meta needs to develop FRONTIER logic models to operate.

With the addition of Bansal and other artificial intelligence researchers, Meta hopes to progress in the artificial intelligence race. This may be difficult to look at this Openai plans to issue an open AI thinking model In the coming weeks – an offer can add more pressure on Meta Open AI’s offers.

This story was updated after publishing in more details.