Meta is reportedly striving to release a new Llama model by the end of the year

Zhitong
2025.08.29 04:00
portai
I'm PortAI, I can summarize articles.

Meta plans to release its next-generation artificial intelligence model Llama 4.X or Llama 4.5 by the end of this year. This model is one of the first projects of Meta's Superintelligence Lab, which is led by former Scale AI CEO Alexander Wang. Meta has restructured its AI department to focus on training, research, products, and infrastructure. Zuckerberg stated that the company has made good progress on Llama 4.1 and 4.2 and is committed to developing more advanced AI models

According to foreign media reports, Meta Platforms (META.US) plans to release its next-generation artificial intelligence model by the end of this year, internally referred to as Llama 4.X or Llama 4.5. Insiders say this is expected to be one of the first projects of the Meta Superintelligence Labs.

The Meta Superintelligence Labs is led by Alexandr Wang, the former CEO of Scale AI, who was hired after Meta invested $14.3 billion in Scale AI to acquire a 49% stake. It is understood that Meta has reorganized the Meta Superintelligence Labs into four teams, each focusing on training, research, product, and infrastructure work.

Meta first announced the Llama 4 series models in April, which include Maverick, Scout, and Behemoth. Scout and Maverick were released on April 5, while Behemoth, originally scheduled for release in June, is still under development.

Reportedly, Meta CEO Mark Zuckerberg is investing heavily in building new data centers and attracting top AI researchers with compensation packages worth hundreds of millions of dollars. Recently, Meta reorganized its internal AI department and named it the "Meta Superintelligence Labs," aiming to develop human-level AI capabilities and apply this technology across all of the company's products.

In July of this year, Zuckerberg stated during the second-quarter earnings call that the establishment of the Meta Superintelligence Labs aims to focus on developing the next-generation AI models and confirmed that "good progress is being made on Llama 4.1 and 4.2," while also working on more cutting-edge next-generation models