Google’s Gemini 2.5 Flash Model Is Now Available
▼ Summary
– Google’s Gemini 2.5 Flash Image AI model, previously known as “nano-banana,” is now generally available for full-scale use after completing its testing phase.
– The model offers 10 aspect ratios across four styles to support content creation for various formats like landscapes and social media posts.
– It is accessible via the Gemini API and Vertex AI, costing $0.039 per image, and includes developer resources to help users get started.
– Gemini 2.5 Flash Image excels at maintaining subject consistency across images, making minor edits, and combining images without common AI errors like extra fingers.
– All images generated or edited with the model contain an invisible Synth ID watermark to identify them as AI-generated, addressing concerns about deepfakes.
Google’s latest image generation model, Gemini 2.5 Flash Image, has officially completed its testing phase and is now available for general use. Previously known under the playful codename “nano-banana,” this advanced AI tool offers creators and businesses a powerful new resource for generating visual content. The model supports ten distinct aspect ratios across four primary styles: landscape, square, portrait, and a flexible option, making it simple to produce material for everything from widescreen cinematic scenes to vertical social media posts.
Developers and users can access Gemini 2.5 Flash Image through the Gemini API on Google AI Studio, while enterprise clients will find it on Vertex AI. A key strength of the model is its ability to maintain subject consistency across multiple images. This means a company could generate pictures of the same product in various settings, or a user could create images of a person or character in different outfits, all while avoiding common AI errors like extra fingers or other visual oddities.
Beyond generating images from scratch, the model excels at performing detailed edits based on written instructions, such as removing an object from a photo or blending several images into one cohesive picture. To help new users get started, Google has released comprehensive developer documentation and a “cookbook” with practical guides. Each image generated costs $0.039.
Before its official launch, the model gained significant attention on platforms like LMArena under its “nano-banana” alias. On the same day Google formally announced it, Adobe revealed that the model would be integrated into its Firefly and Express creative tools.
The rapid advancement of image-generating AI has led to a surge in highly realistic synthetic media, raising concerns about deepfakes and digital authenticity. In the absence of broad federal regulations, tech companies are implementing their own transparency measures. Any image created or edited with Gemini 2.5 Flash receives an invisible SynthID watermark. This digital tag allows specially trained AI systems to identify the content as machine-generated, though it remains undetectable to people.
(Source: ZDNET)
