Key Takeaways:
Meta's new Muse Image model lets users generate AI photos by tagging Instagram accounts, replacing third-party tools like Midjourney as the company pushes into generative AI for consumers and advertisers.
Meta released Muse Image, its first in-house AI image generation model, replacing third-party tools from Midjourney and Black Forest Labs as it competes with OpenAI and Google for creator and advertising revenue. The model, developed by Meta Superintelligence Labs under Alexandr Wang, is available free in Meta AI, Instagram, and WhatsApp, with paid tiers starting at $7.99 a month for users who exceed daily limits.
"Muse Image is agentic — it works with Muse Spark to reason through your prompt, search the web, and plan before it generates," Wang said on Threads. The model lets users @mention Instagram accounts in prompts, pulling public photos to incorporate likenesses into AI-generated images, a feature Meta said users can control through privacy settings.
Internal benchmarks show Muse Image trailing OpenAI's GPT Image 2 but beating Google's Nano Banana 2 in single and multi-image editing tasks, according to Meta. The model powers 30 new AI effects rolling out to Instagram Stories in the US, with broader availability planned across Facebook and Messenger later this year. Meta previously relied on Midjourney and Black Forest Labs for image and video generation in its Meta AI app.
The subscription play
Meta's subscription plans, introduced in May, offer tiered access to AI features. Free users get a daily limit of AI-generated images, after which they must purchase a Meta One subscription or wait for their quota to reset. The pricing mirrors a broader industry shift as AI companies seek to monetize generative tools — OpenAI charges $20 a month for ChatGPT Plus, while Google includes AI features in its Google One plans.
The subscription revenue is small relative to Meta's $160 billion annual ad business, but the company is positioning AI tools as a way to deepen engagement and keep users within its ecosystem rather than turning to third-party services like Midjourney or OpenAI's DALL-E.
Advertiser tools and competitive stakes
Muse Image also powers image-generation features within Meta's Advantage Plus advertising service, letting brands create ad variations automatically. Meta said it has been working with businesses and advertisers during the model's development, and that advertisers will see image variants powered by Muse Image in the coming weeks.
The move reduces Meta's dependence on third-party AI vendors and gives it more control over costs and capabilities. Meta spent heavily on AI infrastructure in recent years, with capital expenditures exceeding $35 billion annually, and in-house models reduce the per-inference cost of serving AI features to its 3 billion-plus monthly active users.
OpenAI and Google got a head start in image generation — OpenAI's GPT Image 2 and Google's Nano Banana, released last fall, both gained significant consumer traction. Meta's internal benchmarks suggest it has closed part of the gap, though it still trails OpenAI on quality metrics. The company plans to release Muse Video, a video generation model, at a later date.
Meta shares have gained 18% this year through Monday's close, outperforming the Nasdaq Composite's 12% gain. The company trades at 23 times forward earnings, a discount to Microsoft at 31 times but a premium to Alphabet at 20 times, according to data compiled by Bloomberg. Investors will watch whether AI features can drive measurable engagement or subscription revenue in the coming quarters.
This article is for informational purposes only and does not constitute investment advice.
