Meta launches Muse Image model with Instagram account prompts and QR code generation
Meta has launched Muse Image, the first AI image generation model from Meta Superintelligence Labs, now available in the US through Meta AI app, Instagram, and WhatsApp. The model accepts Instagram accounts as prompts to incorporate users' likenesses and claims to generate functional QR codes with legible styled text.
Meta launches Muse Image model with Instagram account prompts and QR code generation
Meta has released Muse Image, the first AI image generation model from Meta Superintelligence Labs, available now in the US through the Meta AI app, Instagram, and WhatsApp. The model follows the April 2026 launch of Muse Spark, Meta's first entry in its new family of AI models.
Key capabilities
Muse Image accepts Instagram accounts as prompts, allowing users to tag friends' accounts to incorporate their likenesses in generated images. The feature uses the same privacy controls already in place for photo reuse on Instagram. According to Meta, the model "uses advanced reasoning to understand complex prompts, seamlessly blending multiple photos into high-quality creations."
The model claims two unusual capabilities: generating functional QR codes and producing legible text that matches the style of generated images. Meta has not provided benchmark scores or technical specifications for the model.
Integration across Meta products
Muse Image powers over 30 new effects for Instagram Stories, including a disposable camera effect and one called "Puffer." A new preview interface in Instagram's image editor lets users view effects before applying them.
In the Meta AI app, users can draw or write on photos to indicate desired changes. The model connects to Meta's broader ecosystem—users can photograph a room and ask Meta AI to redecorate, pulling furniture from Facebook Marketplace and web listings.
The same image generation and editing features are available in WhatsApp through Meta AI chats.
Pricing and availability
Muse Image is free for "everyday creation," according to Meta, with usage limits that vary by service and location. After reaching the limit, users are prompted to subscribe to Meta One for additional generations. Meta did not disclose specific image limits or Meta One pricing.
The model is currently US-only, with expansion to additional countries and Meta products planned. Meta says a Muse Video model is in active development.
What this means
Meta is positioning Muse Image as a social-first AI model, integrating directly with Instagram's social graph rather than treating image generation as a standalone tool. The Instagram account prompting feature appears designed to replicate the viral interest OpenAI generated with its Sora app's friend video feature before that app was discontinued. Whether users embrace having their likenesses used in AI-generated images—even with existing privacy controls—remains to be seen. The claimed QR code and legible text generation capabilities, if accurate, would address common weak points in image generation models.
Related Articles
Kuaishou launches Kling Image 4.0 with native 8K image generation and up to 10 images per prompt
Kuaishou launched Kling Image 4.0 on Tuesday night, adding native 8K image generation and up to 10 coordinated images from a single prompt. The model is available through an API for developers and enterprise customers.
Thinking Machines Lab releases Inkling 2 Mini: 26B-parameter looped MoE with 3B active and 1M-token context
Thinking Machines Lab has released Inkling 2 Mini, a multimodal Mixture-of-Experts model with about 26B physical total parameters and 3B active per token. It repeats its middle layers three times, supports a 1M-token context, and ships with a DSpark drafter. Hugging Face reports it leads 32 of 48 benchmarks against four comparable models.
NVIDIA releases Nemotron-Labs-Sigma: 14 open 3B and 8B continuous diffusion language model checkpoints
NVIDIA released Nemotron-Labs-Sigma, a set of 14 open continuous diffusion language model checkpoints at 3B and 8B parameters. According to MarkTechPost, Sigma denoises text in a learned 16-D embedding space, distills to few-step samplers, and decodes up to 4.20x faster than autoregressive decoding.
Tencent releases Hy4 flagship weights: 770B-parameter MoE, 1,048,576-token context, conditional license
Tencent released downloadable weights for its flagship Hy4 on November 20, with a System Card. The multimodal MoE has a 770-billion-parameter language backbone, 49 billion active parameters per text token and a 1,048,576-token context window, under a license with commercial conditions for model-as-a-service operators.
Comments
Loading...