Meta hires key OpenAI researchers to accelerate development of proprietary inference models

Meta, the company behind Facebook and Instagram, has developed its own AI app Meta AI and the large-scale language model Llama , among other AI-related developments. It has now been revealed that Meta is poaching key researchers from its rival OpenAI.

Meta hires key OpenAI researcher to work on AI reasoning models | TechCrunch
https://techcrunch.com/2025/06/26/meta-hires-key-openai-researcher-to-work-on-ai-reasoning-models/

Meta Taps Top Researchers From Google, Sesame for New AI Lab - Bloomberg
https://www.bloomberg.com/news/articles/2025-06-11/meta-hires-top-researchers-from-google-sesame-for-new-ai-lab

According to information obtained exclusively by tech media TechCrunch from sources, Meta has poached OpenAI's leading researcher, Trapit Bansal, to help accelerate the development of inference models for its new Superintelligence Lab.

OpenAI spokesperson Kayla Wood confirmed to TechCrunch that Bansal had left the company, and according to his LinkedIn profile , he will leave OpenAI in June 2025.

Bansal has worked at OpenAI since 2022 and is known for playing an important role in launching a reinforcement learning research project with the company's co-founder Ilya Satkivar. Bansal made a significant contribution to developing the basis of OpenAI's first inference model, o1.

OpenAI announces 'OpenAI o1' and 'OpenAI o1-mini' AI models with complex inference capabilities, demonstrating high capabilities in programming and mathematics - GIGAZINE

The Meta Superintelligence Lab, which Wood will join, is planned to be built as part of the company's investment in data labeling service Scale AI. Meta's Superintelligence Lab construction plan is summarized in the following article.

Meta plans to build an AI lab led by Scale AI CEO Alexander Wang specializing in the pursuit of 'superintelligence,' and offers hundreds of millions to tens of billions of yen in compensation to lure AI researchers from OpenAI and Google - GIGAZINE

TechCrunch points out that the addition of Bansal to the Superintelligence Lab could enable Meta to develop state-of-the-art inference models that can compete with industry-leading inference models such as OpenAI's o3 and DeepSeek's R1. At the time of writing, Meta has not released its own inference models.

Meta CEO Mark Zuckerberg has been actively recruiting in recent months to build a superintelligence lab. Some of the best AI researchers joining Meta are said to have been offered compensation packages worth $100 million. However, it is unclear how much Bansal was offered in this latest poaching.

According to Bloomberg, Bansal will join former Google DeepMind researcher Jack Ray and former Sesame machine learning researcher Johan Schallwig at Meta, along with former OpenAI AI researchers Lucas Bayer, Alexander Kolesnikov and Xiaofua Zai, who will also join Meta's Superintelligence Lab.

On the other hand, it has been reported that Meta has been losing the original developers of LLM and Llama, which are core products in the AI field, to competitors one after another.

Meta has lost 11 of the 14 original Llama development team members to competitors - GIGAZINE

Meta has also reportedly been in acquisition talks with AI startups such as erplexity AI and Thinking Machines Lab.

Meta is in acquisition talks with a number of AI startups, including Perplexity AI and Thinking Machines Lab - GIGAZINE