Meta Platforms is rolling out Muse Image, its first image-generation model from Meta Superintelligence Labs, as it expands generative AI features across its apps.
The model is being integrated into Meta AI, the company’s chatbot. It can generate images from text prompts, use photos as inputs, and support edits through sketches or annotations, according to the company.
Muse Image can also adjust image elements, change styles, and create variations. Meta said the model works with Muse Spark, its text-and-reasoning model, to process prompts before generating images.
The model will power more than 30 new AI effects for Instagram Stories. Meta said it will also enable image generation in direct chats with Meta AI on WhatsApp, starting in limited countries.
Meta plans to bring Muse Image to more countries and to other parts of its apps, including Facebook and Messenger. The company said additional surfaces on Instagram and WhatsApp will also be added later.
Meta reported 3.56 billion family daily active people on average in March 2026, up 4% from a year earlier.
Muse Image includes presets and direct editing tools. Meta said users can restore old photos, test different hairstyles, rework image styles, redesign rooms, render text inside images, and request changes by circling or sketching on generated images.
The company said the room-design feature can use products from the web or Facebook Marketplace. Meta also said Muse Image can render text inside images, including for guides and visual designs.
Muse Image will allow users to @mention Instagram accounts in Meta AI prompts. Meta said this lets the system use public photos from the tagged account to create an image.
The company said users can control how their content is tagged for AI creation. Users who do not want others to reuse or remix their public Instagram content for AI images can turn off the feature through Instagram’s settings.
Basic access to Muse Image through Meta AI will be free. Meta said additional creation capacity will be available through its subscription plans.
Meta also plans to make Muse Image available to advertisers and agencies through Advantage+ creative. The feature will place image generation inside Meta’s existing advertising tools.
In the first quarter of 2026, Meta said ad impressions across its Family of Apps rose 19% year over year, while the average price per ad increased 12%.
Images created with Muse Image in the Meta AI app and on meta.ai will carry Content Seal, Meta’s invisible watermarking system. Meta said the hidden signal is designed to remain intact after cropping, compression, resizing, or screenshots.
Meta is also previewing a detection tool that can check whether an image carries a Content Seal watermark. The company said it plans to extend Content Seal to video later.
Meta said in its AI research blog that Muse Image ranked No. 2 on Arena for text-to-image generation, single-image editing, and multi-image editing, based on human-preference Elo rankings as of July 5, 2026.
The release follows the April launch of Muse Spark, the first model in Meta’s Muse family from Meta Superintelligence Labs. Muse Spark currently powers the Meta AI app and website, and Meta has said it will be rolled out across WhatsApp, Instagram, Facebook, Messenger, and its AI glasses.
Reuters reported in June that Meta had delayed plans to release the Muse Spark API to developers, a detail related to Meta’s broader developer-access plans rather than the Muse Image launch. A Meta spokesperson said at the time that the company was testing the API with early partners and expected to release it that month.
Meta also announced an early preview of Muse Video, its video-generation model. Meta said the model is built on the same pretraining base as Muse Image, supports native audio, and is expected to come to creators and Meta AI.
The company said Muse Video is being evaluated on prompt adherence, visual fidelity, and temporal consistency. Meta also said it is still working on areas such as audio-video synchronisation and physically accurate fast motion.
Meta said in its AI research blog that Muse Video ranked No. 3 on Arena for text-to-video, based on human-preference Elo rankings as of July 5, 2026.
Bloomberg reported that Meta has been developing a cloud infrastructure business that would sell access to AI computing power and hosted AI models. The reported plan could allow developers to access Meta-hosted models and pay for the computing power needed to run them, though Bloomberg said the plans remain in development and could change.
In April, Meta raised its 2026 capital expenditure outlook to $125 billion to $145 billion, citing higher component pricing and additional data centre costs to support future capacity.
Meta has previously used several image-generation tools, including its own Emu model. It has also licensed image-generation technology from companies such as Midjourney.
Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events, click here for more information.
Tech Wire Asia is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。