Meta on Wednesday launched an artificial intelligence model, Muse Spark, it promotes as smarter and faster than what it provided earlier than shaking up its Superintelligence Labs unit.
“Over the last 9 months, Meta Superintelligence Labs recreated our AI stack from the ground up,” the tech titan stated that in a blog post.
Muse Spark succeeds Llama 4, launched by the Silicon Valley-primarily based company a year ago, and could power Meta’s AI app and smart glasses along side Facebook, Instagram, WhatsApp and Messenger features.
For now, Muse Spark is only available in the US.
The latest AI model was defined as being small and rapid by design, able to reasoning by complex questions in science, math and health.
It is the first in a latest Muse series, with the next generation already in expansion.
Llama 4 was behind in the fierce AI race as heavyweight rivals from China, France, and US developed advanced models at a rapid-fire pace.
That prompted Meta chief executive Mark Zuckerberg to overtake its AI team, which saw the departure of its research boss Yann LeCun.
LeCun spent 12 years managing the AI lab at Meta, where Zuckerberg has made the quest for “superintelligence” a priority.
Zuckerberg embarked on a chief recruitment campaign last year to obtain skills for Meta’s efforts, poaching Scale AI co-founder Alexandr Wang and placing him in charge of a newly designed unit called Superintelligence Labs.
Zuckerberg later recruited executives from rivals OpenAI, Anthropic and Google—frequently personally and at heady costs.
In doing so, the tech tycoon broke with the corporation’s previous method of prioritizing development of free, open-access AI models which include Llama.
“The future of Meta AI is rooted in the relationships and context already at the center of your life,” the corporation stated.
“We are constructing towards personal superintelligence—an AI that does not just answer your questions however clearly understands your world due to the fact it’s built on it.”
