Note: Grok Imagine Image 2.0 is xAI's latest stills model on fal. Learn more on the fal Grok Imagine page and in the xAI Imagine docs.
We've added Imagine Image 2.0 for both generation and editing. Create images from a prompt, or refine existing images with natural language — 1K or 2K output, low or medium quality, and up to three input images on edit.
Imagine Image (1.0) stays in the picker so existing billed generations keep their original model id.
Why This Matters
- Sharper stills: Choose 1K or 2K and low or medium quality to match the job — ideation stays cheap, hero frames can go 2K.
- Better edits: Describe the change and attach up to three total input images (source + 2 extras). Default aspect is auto so the first input's framing is preserved.
- Same workflow: Available in Image Studio generate and edit, next to Imagine Image 1.0.
What's New
- New image model: Imagine Image 2.0 (generate + edit).
- Resolution & quality: 1K or 2K, low or medium (default 1K + medium).
- Edit references: Up to 3 input images. 1.0 edit is unchanged (single
image_url). - Aspect ratios: Wide fal set including 20:9, 19.5:9, and 9:16. Edit also supports auto.
How To Use It
Select Imagine Image 2.0 in the image generation or edit model picker. For generation, write a prompt and pick aspect ratio, 1K/2K, and quality. For editing, add a source image (and optional references) and describe the change.