Meta Platforms Inc. (NASDAQ:META) announced the launch of its first major artificial intelligence model on Wednesday, nine months after the high-profile hiring of Scale AI’s Alexandr Wang.

The AI model, called Muse Spark, initially code-named Avocado, is the first in the new Muse series. Meta Superintelligence Labs has rebuilt the AI stack from scratch in the past nine months, marking a faster development cycle than any previous ones.

Muse Spark Powers AI, Shopping, And API

Meta is positioning Muse Spark not as a top-tier model but is highlighting its efficiency and competitive performance across various tasks. The company plans to open-source future versions of the model, a shift from its current proprietary stance.

"Muse Spark offers competitive performance in multimodal perception, reasoning, health, and agentic tasks," Meta said.

The Mark Zuckerberg-led company is aiming to create a new revenue stream by providing third-party developers with access to Muse Spark's core technology through an API. This model currently powers the company's digital assistant on both the standalone Meta AI app and its desktop website.

Muse Spark will include a Shopping mode that helps users buy clothes or decorate rooms by leveraging inspiration and ideas from creators and communities on its apps.

The new model will soon be integrated into Facebook, Instagram, WhatsApp, Messenger, and the company’s Ray-Ban Meta AI glasses. It is also expected to power the company’s Vibes AI video feature in the Meta AI app.

Mizuho Securities sees Meta's early Muse Spark launch as a positive sign of its competitiveness in advanced AI. It highlights that boosting usage, especially via Shopping mode, could drive “significant” monetization for the company through search and ad targeting, and that the pre-earnings release strengthens Meta's earnings outlook, as reported by Investing.com.

Meta Doubles Down On AI Despite Delays

The launch of Muse Spark was reportedly delayed due to underwhelming performance in tests against Google’s (NASDAQ:GOOGL) (NASDAQ:GOOG) Gemini and Anthropic AI models. Due to the performance gap, Meta had postponed Avocado's launch to at least May from its earlier planned March release.

Meta’s commitment to AI development, particularly superintelligence, is evident in its significant infrastructure expansion, with 2026 capital expenditure projected to reach a record-high range of $115 billion to $135 billion. The savings from reduced employee compensation are reportedly being funnelled directly into Meta Superintelligence Labs, a core initiative focused on developing AI that can outperform human capabilities.

Alexandr Wang outlined the company’s vision on superintelligence at the India AI Impact Summit 2026. He described it as AI that understands a user’s goals, interests, and daily routines, helping them pursue healthier habits, complete projects, and deepen relationships.


Benzinga's Edge Rankings place Meta in the 89th percentile for quality and the 84th percentile for growth, reflecting its strong performance in both areas. Benzinga’s screener allows you to compare Meta’s performance with its peers.

META Price Action: On a year-to-date basis, Meta declined 5.84%, as per Benzinga Pro. On Wednesday, the stock climbed 6.50% to close at $612.42.

Disclaimer: This content was partially produced with the help of AI tools and was reviewed and published by Benzinga editors.

Image via Shutterstock