Meta’s New AI Model Could Arrive by Year’s End
Meta, the parent company of Facebook, is rapidly progressing on its Superintelligence Lab’s first product, with a potential launch later this year. According to a report from Business Insider, a division within the lab, known as TBD (to be determined), is working hard to prepare the new Llama 4.X AI model for release by the end of 2023.
This push for a timely launch comes after previous Llama 4 models, called Scout and Maverick, received mixed reviews when they were introduced in April. Developers found these models lacking in essential capabilities such as coding and following instructions. As a result, the TBD team is not just focused on launching Llama 4.X, but they are also addressing issues with the earlier models.
Meta’s CEO, Mark Zuckerberg, mentioned during a recent earnings call that the lab aims to enhance its AI offerings, noting that progress is being made on not only Llama 4.1 and 4.2 but also the upcoming models that will explore new boundaries in AI technology.
Since the formation of the Superintelligence Lab in June, Meta has aggressively recruited top AI talent, offering generous compensation packages to attract experts from leading companies like OpenAI, Apple, and Google DeepMind. However, the new lab has already encountered challenges, including losing several key personnel shortly after its establishment.
With much at stake, Meta is working diligently to ensure its latest AI model meets expectations and helps the company strengthen its position in the AI landscape.
