
Meta Chief Executive Mark Zuckerberg has spent billions of dollars over the past year to overhaul the company’s artificial intelligence efforts and assemble an elite team of researchers, in a bid to push the social networking giant back into the fast-moving AI race.
The first results of that strategy are now emerging with the launch of Muse Spark, the first model built by Meta Superintelligence Labs, the firm’s revamped AI unit headed by Alexandr Wang.
Shares of the company extended gains to trade up nearly 7%.
According to Reuters, the stakes are especially high for Meta after it hired Scale AI CEO Alexandr Wang last year under a $14.3 billion deal and offered some engineers pay packages of hundreds of millions of dollars to staff a new superintelligence team.
U.S. tech giants are under pressure to prove their massive AI outlays will pay off.
In a departure from Meta’s prior open-source strategy, Muse Spark is a closed model, meaning its code will not be publicly available.
03 Apr 2026 - Vol 04 | Issue 65
The War on Energy Security
The company said it plans to offer the model in private preview to select partners through an API and may open-source future versions.
Superintelligence refers to AI machines that could outthink humans.
The performance of the new model will be closely watched by investors and industry observers to assess whether it can help Meta narrow the gap with leading AI players such as Google, OpenAI and Anthropic, which continue to release powerful models at a rapid pace.
Anthropic said on April 8 that its latest AI model, Mythos, is so powerful that it will not be publicly released.
Muse Spark is the first in a new series of models from the team and is part of a family of models known internally as Avocado.
The model will initially be available only on the lightly used Meta AI app and website and will, in the coming weeks, replace the existing Llama models powering chatbots on WhatsApp, Instagram, Facebook and Meta's collection of smart glasses.
Meta noted that while Muse Spark is a “powerful foundation,” it remains an “early data point” in the company’s trajectory, with several larger models still in development.
"This initial model is small and fast by design, yet capable enough to reason through complex questions in science, math, and health. It is a powerful foundation, and the next generation is already in development," the company said in a blog post.
In January, Zuckerberg said on the company’s earnings call that he “expects the first models will be good, but more importantly, we will show the rapid trajectory that we’re on.”
“Then, I expect us to steadily push the frontier over the course of the year as we continue to release new models,” he added.