Overview
- Hands-on reviews report quick, text-driven tweaks that preserve a subject’s identity across multi-turn edits and merged photos.
- The capability is available in the Gemini app for consumers and through the Gemini API, Google AI Studio, and Vertex AI for developers and enterprises.
- Testers note practical gaps including the model’s inability to perform precise crops to specific aspect ratios and occasional facial or fine-text artifacts.
- Google adds a visible AI mark and an invisible SynthID watermark to edited images, though the SynthID detector remains limited in availability and visible marks can be cropped.
- The model sits atop LMArena’s image-editing leaderboard as users showcase use cases like background replacement, multi-image composites, and interior reimagining.