Meta is doubling down on generative AI. On July 7, 2026, the company announced the rollout of Muse Image, its first dedicated image generation model, built by Meta Superintelligence Labs the AI division formed last year under Chief AI Officer Alexandr Wang. It marks the second major release from the lab, following the Muse Spark language model that debuted in April.
What Muse Image Brings to the Table
Unlike a typical text-to-image generator, Muse Image is designed to handle complex, multi-step prompts and accept photos as inputs. Users can edit generated images directly by sketching or annotating on them, making the creative process more interactive than a one-shot prompt-and-render workflow. Meta says the model uses visual reasoning to adjust individual elements, swap styles, and produce refined variations without needing constant re-prompting.
One notable and slightly controversial feature: users can tag public Instagram accounts and pull in photos from those profiles when generating images. The option is enabled by default, though people can opt out of being tagged or having their photos used this way.
Where You’ll See It
Muse Image is rolling out across Meta’s ecosystem in phases:
- Instagram Stories gets more than 30 new AI-powered effects.
- WhatsApp users can generate images directly within chats with Meta AI, starting in select countries.
- Facebook and Messenger integration is planned for later this year.
- Meta AI Shopping will use the model for a room-restyling feature, letting shoppers preview catalog products in their own spaces.
Free Access, Paid Upside
Basic image generation through Meta AI remains free. But heavier users and creators who want expanded capabilities will need one of Meta’s monthly subscription plans, which launched back in May. Hit your free limit, and you’ll either wait for a reset or pay up.
On the business side, Muse Image is being folded into Meta’s Advantage+ advertising suite, where more than 8 million advertisers already use some form of generative AI creative tools. Early testers reportedly saw improvements in photorealism and product accuracy in their ad creatives.
How It Stacks Up
Meta has previously leaned on third-party models from Midjourney and Black Forest Labs for its image features, while rivals like Google and OpenAI got a head start in this space. Google’s Nano Banana became a breakout hit last fall. According to Meta’s own benchmarks, Muse Image outperforms Google’s Nano Banana 2 in several editing tasks but still trails OpenAI’s GPT Image 2.
The Bigger Picture
The launch comes as Meta pushes hard to close the AI gap with competitors including reports of internal friction over the pace of development and controversial staffing shifts. Meta also teased Muse Video, an early-preview video-generation model, suggesting the Muse family is set to expand further. Whether Muse Image opens up to outside developers remains an open question, but for now, it’s a clear signal that Meta wants generative visuals baked into every corner of its apps.
One thought on “Meta Expands Its Generative AI Push With Muse Image Rollout”