Meta has reportedly postponed the launch of its highly anticipated Behemoth AI model due to ongoing performance concerns. Initially set for release in April, the company pushed the date to June before announcing that the rollout would now be delayed until at least the fall, according to a report by The Wall Street Journal on May 15. Meta has yet to provide a public timeline for the product’s release.
The delay is reportedly caused by challenges in improving the AI model, with internal concerns that Behemoth may not meet the performance standards previously claimed by the company. In public statements, Meta had boasted that Behemoth outperformed similar models from companies like Anthropic, Google, and OpenAI on certain tests.
In addition to the Behemoth delay, Meta has also been focusing on its other AI efforts. The company recently unveiled its latest Llama AI models, including Llama 4 Behemoth, positioning it as one of the most powerful models yet, designed to act as a mentor for new AI models. CEO Mark Zuckerberg announced during the company’s first-quarter earnings call that Meta would be increasing investments in AI data centers to meet the growing demand for computing power, with capital expenditures expected to reach $64 to $72 billion this year.
Despite the setbacks, Meta’s open-source AI models, Llama 4 Scout and Llama 4 Maverick, are gaining attention as more affordable alternatives to competing closed-source models, which could help businesses reduce AI deployment costs. Meta’s continued AI efforts signal the company’s commitment to leading in the AI space.
