Meta launches Muse Spark
Meta introduced Muse Spark, billed as an early step toward 'personal superintelligence,' and the model helped push Meta’s AI app up the App Store charts. The release is a reminder that competitor AI offerings are moving fast and gaining visibility in consumer channels. (x.com) (x.com)
Meta’s new artificial intelligence model was released on April 8, and within about a day the Meta AI app jumped from No. 57 to No. 5 on the United States Apple App Store. Appfigures also showed a sharp rise in installs right after the launch. (techcrunch.com) Sensor Tower estimated roughly 46,000 United States iPhone downloads on April 8 alone, up 87% from the day before. In Canada, the United Kingdom, France, and Germany, iPhone downloads rose 51%, 32%, 27%, and 25% day over day. (techcrunch.com) The model is called Muse Spark, and Meta says it is the first model from Meta Superintelligence Labs, the new group the company built after reshuffling its artificial intelligence effort. Meta says Muse Spark already powers the Meta AI app and the Meta AI website. (about.fb.com) Meta is not pitching this one as its biggest brute-force system. In its own announcement, the company said Muse Spark is “small and fast by design,” which is a clue that it wants a model cheap enough to run inside products people open every day. (about.fb.com) (cnbc.com) Meta says the model can work on text, images, science questions, math questions, and health questions, and it can split one problem across multiple subagents at the same time. Meta compared that to planning a Florida family trip by having separate agents build the itinerary, compare Orlando with the Florida Keys, and find child-friendly activities in parallel. (about.fb.com) That “multiple subagents” idea is Meta’s way of buying more thinking time without making one answer feel painfully slow. TechCrunch reported that Meta plans a “Contemplating” mode for harder questions, where several agents collaborate on the same task before the app returns one answer. (techcrunch.com) The backstory is that Meta’s last big artificial intelligence push, the Llama 4 family, did not land the way the company wanted. CNBC reported that the rollout failed to captivate developers, and Meta responded by rebuilding its stack and putting former Scale AI chief executive Alexandr Wang in charge of the new lab. (cnbc.com) (techcrunch.com) Wang came over as part of Meta’s $14.3 billion investment in Scale AI for a 49% stake, according to TechCrunch and CNBC. Muse Spark is the first big public test of whether that expensive reset can turn Meta from a company with strong open models into one with a consumer artificial intelligence product people actually choose to open. (techcrunch.com) (cnbc.com) Meta also says Muse Spark will roll out beyond the standalone app to WhatsApp, Instagram, Facebook, Messenger, and its artificial intelligence glasses in the coming weeks. The company said it will also offer private application programming interface access to selected partners, which means this is both a consumer launch and the start of a developer business. (about.fb.com)
