Mark Zuckerberg

Muse Spark 1.1 is strongest at agentic performance, tool use, and computer use. It does well on long-running tasks with 1M token context window, can delegate execution to sub-agents running in parallel, and is trained to use computer interfaces on desktop, mobile, or browser.

Today we're releasing Muse Spark 1.1 -- a strong agentic and coding model at a very low price. It's available through our new Meta Model API and in Meta AI.

The Meta Model API allows developers to build using Muse Spark for the first time. Our focus is on delivering strong agentic and multimodal models at very low cost. More to come soon.

Our new line of @metaglasses is available today. Three shapes, 26 style combos, with our most advanced Meta AI built in. Plus, three custom styles designed by @kyliejenner.

500M monthly actives on Threads in less than 3 years. Thanks for making this platform what it is. 🙏

Interesting Biohub conversation with No Priors, Priscilla, and Alex Rives. We covered our latest release of a world model for protein biology (ESM) and what it means to do frontier biology alongside frontier AI. 10 years ago we set a goal to help the scientific community cure or prevent all diseases this century. With new AI advances, I'm optimistic we'll be able to get there a lot sooner.  https://open.spotify.com/episode/2bufgHVuFxdBr6vKUkItdg?si=2Ap4WgeGTou8MCwUVlhoaw

America is going to need hundreds of thousands of skilled tradespeople to build the infrastructure for our country to lead in AI.  People need access to the education and opportunity to secure these jobs. We believe the future is for everyone. Today, we're launching America's Workforce Academy with academic and workforce partners to provide free training and direct pathways to jobs building this infrastructure.

We want to deliver personal superintelligence for everyone -- and we want as many people as possible to help build that future with us. Check it out: https://about.fb.com/news/2026/06/americas-workforce-academy-free-skilled-trade-training/

Today Biohub is sharing a major scientific advance: our first world model for protein biology. This system can help scientists discover and design proteins that will hopefully lead to new medical therapies.

This release consists of two AI models and an atlas: ESMC, our state-of-the-art language model that represents proteins; ESMFold2, designed to transform ESMC’s sequence representations into atomically resolved 3D structures; and ESM Atlas, a dataset and tool that makes these representations navigable across 6.8 billion protein sequences. All are openly available at biohub.ai/esm/protein so scientists everywhere can use them and build on them.

Today we're starting to roll out Incognito Chat with Meta AI on WhatsApp and the Meta AI app -- a completely private way to interact with AI, similar to how end-to-end encryption means no one can read your conversations, even Meta or WhatsApp. This is the first major AI product where there is no log of your conversations stored on servers.

Incognito Chat handles all AI inference in a Trusted Execution Environment that ensures your messages are not accessible to us. The conversations on your phone also disappear when you exit the session. This is different from other disappearing AI products where your conversations logs often remain on other companies' servers for many months.