You do not need to pay. Open Gemini and describe the picture; Nano Banana 2 is free there, with a cap on how many you make. Or open ChatGPT on the free plan, which includes limited image creation. Gemini puts an invisible mark in the file saying it was AI-made. The third door is to download an open-weight image model and run it yourself: no cap, nothing sent to a server, but it needs a decent graphics card.
Key facts
- $0for Nano Banana 2 in the Gemini app, with higher limits on the paid plans
- $0for ChatGPT's free plan, listed as limited image creation
- 2marks Gemini puts on every image it makes, one invisible and one you can see
- $19.99a month, Google AI Pro, the step up when the free cap is not enough
- gemini.google image generation overview and gemini.google/subscriptions; chatgpt.com/pricing; all read 6 November 2026
Which free option gives the best images?
For most people, Gemini. Google’s free tier includes its current image model, which it calls Nano Banana 2, and the overview page describes it as free with higher generation limits on the paid plans. It generates, edits a photo you upload, blends several pictures, keeps a character consistent across images and renders text inside a picture, which older free tools could not do. ChatGPT’s free plan also makes images and the quality is good, but OpenAI describes the free tier only as limited, and the better-quality wording starts at the $20 Plus plan.
| Where | Cost | Cap | Mark on the image | Needs |
|---|---|---|---|---|
| Gemini app (Nano Banana 2) | $0 | Yes, unpublished | Invisible SynthID and a visible mark | A Google account |
| ChatGPT free plan | $0 | Yes, unpublished, counted per day | Not stated on the pricing page | An OpenAI account |
| Open-weight model on your PC | $0 | None | None | A graphics card with memory to spare |



