Overview

• Google has confirmed the once-anonymous model as Gemini 2.5 Flash Image and made it available through the Gemini app and Google AI Studio.
• Hands-on tests describe 1–2 second turnarounds and reliable, plain‑language editing that keeps subjects’ identities consistent across changes.
• Real‑world trials show convincing composites such as celebrity selfies, while tougher scenarios can introduce errors like altered accessories or mismatched faces.
• All generated images include a visible AI label and an invisible SynthID watermark to help signal provenance.
• Community posts report free daily use via Gemini and AI Studio and an API priced around $0.039 per image, as third‑party tools such as Banana AI promote offerings built on the same model.