Meta Launches Llama 4 AI Models with Ghibli Art Capabilities:

Meta has unveiled the first two models in its next-gen Llama 4 AI lineup: Llama 4 Scout and Llama 4 Maverick. These models are multimodal, supporting both text and image inputs, and come with a massive 10 million token context window. The Scout model features 17 billion active parameters and runs efficiently on a single NVIDIA H100 GPU. Meanwhile, Maverick takes performance up a notch with a total of 400 billion parameters.

Ghibli-Style Art Generation Now Available with Meta AI
Llama 4 integrated into Meta AI now supports generating artistic images, including those inspired by Studio Ghibli. During testing, the model was able to generate four Ghibli-style images from a simple text prompt like “a girl waiting at the bus stand.” Image generation is fast, although the output quality is slightly below that of ChatGPT.

Limitations: No Photo Uploads for Style Transfer Yet
Unlike ChatGPT and Elon Musk's Grok, Meta AI does not currently support converting user-uploaded photos into Ghibli-style artwork. Its art generation is limited to prompt-based creation, lacking the photo transformation feature available in some other AI tools.

Exclusive Feature: Animate Your Ghibli Art
Where Meta AI shines is in its ability to animate the images it generates. Users can hit the “Animate” button after generating images to turn still Ghibli-style artwork into animations—something not currently available on competing platforms like ChatGPT or Grok.