Mark Zuckerberg’s Meta Platforms is postponing the launch of its leading “Behemoth” AI model because of concerns regarding its capabilities, according to the Wall Street Journal on Thursday, which referred to individuals familiar with the situation.

The company’s engineers are facing challenges significantly enhancing the capabilities of its Behemoth large-language model, leading to internal questions regarding whether the improvements over prior versions are substantial enough to merit a public release, the report noted.

During its early development, Behemoth was initially intended for release in April to coincide with Meta’s first AI conference for developers; however, the internal target for the model’s launch was later moved to June, as reported. The release has now been postponed to fall or later.

The social media giant did not immediately respond to a Reuters request for comment.

Meta had said in April it was previewing Llama 4 Behemoth, which it called “one of the smartest LLMs in the world and our most powerful yet to serve as a teacher for our new models.”

It released the latest version of its LLM Llama, called the Llama 4 Scout and Llama 4 Maverick, that month.