Top News

Meta unveils new AI model as it tries to catch up with Google and OpenAI after spending billions

Samira Vishwas | April 9, 2026 7:24 PM CST

Meta on Thursday unveiled a new artificial intelligence model called Muse Spark after hiring Scale AI’s Alexander Wang, as the company looks to become more competitive in the AI space dominated by the likes of Google, OpenAI, and Anthropic. It is the first AI model to be launched under the new AI division that Mark Zuckerberg spent billions building last year.

“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,” Meta said in a blog post. “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.”

Called Muse Spark and originally code-named Avocado, the new AI model outperforms Meta’s previous AI models in writing and reasoning and comes close to top competitors like Google, OpenAI, and Anthropic, but still lags in coding. The announcement comes a day after Anthropic’s latest model, Mythos, was deemed too powerful to release due to cybersecurity risks.

Meta’s new AI model puts the company back in the race, though industry insiders will be closely watching how the market reacts to Muse Spark. Google’s Gemini, OpenAI’s GPT, and Anthropic’s Claude currently lead the development of foundational AI models.

Meta has spent billions to create a highly specialized AI team to form a “superintelligence” lab last year under the leadership of Alexander Wang, the 29-year-old Silicon Valley entrepreneur whom Zuckerberg hired as chief AI officer.

Although Meta is committed to investing $600 billion and has, in recent months, used generative AI to improve its advertising business and set up new data centers to compete in the AI race, the company is nowhere near its rivals in capturing the market. OpenAI and Anthropic are now collectively valued at over $1 trillion, and Google’s Gemini technology and services have gaineda significant traction, particularly in the consumer market.

Muse Spark will be a proprietary model, though Meta says it may open-source future versions, unlike its earlier Llama models. The company also claims improved training methods and infrastructure now allow it to build smaller AI models that match the performance of older mid-sized models while using significantly less computing power.

Story continues below this ad

Muse Spark will be available through Meta’s standalone AI app and will be coming to WhatsApp, Instagram, and its AI smart glasses in the coming weeks. The AI model has been unveiled after a series of internal delays. For a change, Meta will be experimenting with a new AI model revenue stream by offering third-party developers access to Muse Spark’s underlying technology via an API.

Anuj Bhatia is a seasoned personal technology writer at indianexpress.com with a career spanning over a decade. Active in the domain since 2011, he has established himself as a distinct voice in tech journalism, specializing in long-form narratives that bridge the gap between complex innovation and consumer lifestyle.
Experience & Career: Anuj has been a key contributor to The Indian Express since late 2016. Prior to his current tenure, he served as a Senior Tech Writer at My Mobile magazine and held a role as a reviewer and tech writer at Gizbot. His professional trajectory reflects a rigorous commitment to technology reporting, backed by a postgraduate degree from Banaras Hindu University.
Expertise & Focus Areas: Anuj’s reporting covers the spectrum of personal technology, characterized by a unique blend of modern analysis and historical context. His key focus areas include:

Core Technology: Comprehensive coverage of smartphones, personal computers, apps, and lifestyle tech.

Deep-Dive Narratives: Specializes in composing longer-form feature articles and explainers that explore the intersection of history, technology, and popular culture.

Global & Local Scope: Reports extensively on major international product launches from industry titans like Apple and Google, while simultaneously covering the ecosystem of indie and home-grown tech startups.

Niche Interests: A dedicated focus on vintage technology and retro gaming, offering readers a nostalgic yet analytical perspective on the evolution of tech.