Details
The tool is built on the NVIDIA Edify model architecture (part of NVIDIA Picasso) and also powered by Bria, a partner AI platform. Customers enter text prompts of up to 100 words, and the model generates four unique image variations per prompt. An updated model launched in July 2024 doubled generation speed to roughly six seconds for four images and added 4K upscaling, longer prompt support (up to 250 words), and advanced camera controls. The model includes filters to block prompts that could generate problematic content, and generated images include watermarks identifying them as AI-created. Images are not added to Getty's licensable libraries for others to license.
Have evidence about Getty Images's AI practices? Submit a report.
Report a Sighting →