Glossary
Upscaling
Also known as: AI upscale, super-resolution
Increasing the resolution of an existing image or video using an AI model. Adds detail by inferring what the higher-resolution version would look like, rather than just interpolating pixels.
In depth
What it means in practice
AI upscaling is the post-generation step that takes a 1024×1024 image and produces a 4K version with detail that was not in the original. It is not a perfect substitute for generating at native high resolution but is much faster and cheaper for cases where you decide after the fact that you need a larger output.
Upscaling is a separate operation on hilens (with its own credit cost) rather than a free post-process. For workflows that consistently need high resolution, generating at native 4K (e.g., Nano Banana 2) is more efficient than generating at 1024 and upscaling.
Related
Output resolution
The pixel dimensions of a generated image or video. Common values: 1024×1024 (default for most image models), 4K (Nano Banana 2 portrait), 1080p (most video models).
Inpainting
Replacing or editing a specific region of an image while keeping the rest unchanged. The user masks the region; the model generates new content matching the surrounding context.
Video generation
AI-driven creation of video clips from text prompts, reference images, or both. Output is typically a short clip (5-10 seconds) at fixed resolution.