Meta has unveiled Muse Spark, its first major artificial intelligence model under a new internal unit, as the social media giant makes a concerted push to reestablish itself as a serious force in a generative AI race it has so far struggled to win.
The model is the first to emerge from Meta Superintelligence Labs, an AI division stood up under the leadership of Alexandr Wang — the former CEO of Scale AI who joined Meta nine months ago as part of the company’s landmark $14.3 billion investment in the data labelling firm. Wang now serves as Meta’s chief AI officer, and Muse Spark represents his most visible deliverable since taking the helm.
Originally developed under the internal codename Avocado, Muse Spark is the first in a new Muse series of models — a deliberate break from the Llama lineage that has defined Meta’s AI identity in recent years.
“Over the last nine months, Meta Superintelligence Labs rebuilt our AI stack from the ground up, moving faster than any development cycle we have run before,” the company said in a blog post on Wednesday.
The post described Muse Spark as intentionally compact and fast, yet capable of reasoning through complex problems in science, mathematics, and health — framing it as a powerful foundation with a next-generation successor already in development.
The launch carries clear redemptive intent. Meta’s previous flagship offering, the Llama 4 family of models, underwhelmed at debut — a setback that left the company trailing rivals who have set a blistering pace in the AI development race. Muse Spark is Meta’s attempt to reset that narrative, with the company emphasising the model’s efficiency and competitive performance across a range of tasks.
To monetise the technology beyond its own platforms, Meta is also exploring a new revenue stream: opening Muse Spark’s underlying architecture to third-party developers via an API — a move that would place it in more direct commercial competition with OpenAI, Anthropic, and Google.
The gap Meta is trying to close is not small. OpenAI and Anthropic are now collectively valued at over $1 trillion, and Google’s Gemini suite has gained meaningful traction — particularly among consumers. While Meta has successfully deployed generative AI to sharpen its advertising business and drive internal efficiencies, it has yet to make a significant dent in the AI model market itself, even as its rivals have accelerated.
The commercial prize at stake makes the urgency understandable. The global generative AI market is projected to expand at more than 40 percent annually, scaling from roughly $22 billion in 2025 to nearly $325 billion by 2033, according to Grand View Research.
Serena Williams Is Building A New Legacy
Investors appeared encouraged by the announcement: Meta’s stock surged almost 9 percent on April 8, its sharpest single-day rally since January — though the broader market also rose sharply that day after President Donald Trump announced a two-week suspension of Iran-related military action, sending oil prices tumbling and lifting equities across the board.
Whether Muse Spark is the breakthrough that closes the competitive distance — or merely a promising step in a much longer race — will likely become clearer as the model faces real-world scrutiny in the weeks ahead.
Taken together, the infrastructure spend, the platform-wide deployment, and the commercial API ambitions sketch the outline of a company not simply iterating on AI — but attempting a wholesale repositioning within it.
The new Muse Spark will be proprietary, with the company saying there is “hope to open-source future versions of the model.” The company had been taking an open-source approach to AI with its Llama family of models.
Apple Buys Polish Startup MotionVFX In A Big Push To Rival Adobe
Meta said in a technical blog about the new model that improved AI training techniques along with rebuilt technology infrastructure has enabled the company to create smaller AI models that are as capable as its older midsize Llama 4 variant for “an order of magnitude less compute.”
Meta is also experimenting with a new AI model revenue stream by offering third-party developers access to Muse Spark’s underlying technology via an API. Currently, only unspecified “select partners” can access the AI model’s “private API preview,” Meta said, but it plans to eventually offer paid API access to a wider audience at a later date.
The new model now powers the company’s digital assistant in the standalone Meta AI app and desktop website, Meta said. Muse Spark will debut in the coming weeks inside Facebook, Instagram, WhatsApp and Messenger, as well as in the company’s Ray-Ban Meta AI glasses, it said. Meta also plans for Muse Spark to eventually power the company’s Vibes AI video feature in the Meta AI app. That service currently uses AI models from third parties such as Black Forest Labs, Meta said.
The company said the revamped Meta AI with Muse Spark will also contain a Shopping mode that will be able to help people buy clothes or decorate rooms.