Detection
Can it detect images made by Gemini, Nano Banana or ChatGPT?
Often, yes. OpenAI images are verified through OpenAI's endpoint; Google images through their Content Credentials signature when it survives, otherwise through the classifier.
Updated
ChatGPT, gpt-image, DALL-E
Every image OpenAI has issued since May 2026 carries a SynthID watermark and a Content Credentials manifest, and OpenAI runs a public verification endpoint. Our API sends the image file there and reports the result under Image watermark. This works even when a platform stripped the metadata, because the watermark is in the pixels and OpenAI can read its own. See The OpenAI provenance check.
Gemini, Imagen, Nano Banana, Pixel
Google signs its generated media with Content Credentials as Google LLC. When the platform preserved the metadata, Slop Squash validates the signature and names the generator outright; that is a Strong signal at 0.97. See Content Credentials.
When the metadata is gone, which is the common case on X, Reddit, Instagram and Facebook, Google's own SynthID watermark is still in the image but only Google can read it. There the generator classifier (Pro triggers it; everyone sees the cached result) gives a probability and attributes the image to Nano Banana, Imagen or another generator. See SynthID: what can and cannot be checked.
Midjourney, Stable Diffusion, Flux, Ideogram, Leonardo
Usually no signature. Metadata tags name these generators when preserved; otherwise it is the classifier.
Text from any model
No watermark anyone else can read, so text is scored on writing patterns. See Can it detect AI-written text?.