Gemini 2.0 Flash native image generation (experimental)
Gemini 2.0 Flash gets experimental native image output in AI Studio and the API, moving image generation and editing inside the multimodal LLM.
Image generation shifts from a separate Imagen model into the chat model itself, enabling edit-by-conversation. Trusted testers had native image output since December 2024; a preview image-generation model followed 2025-05-07; Gemini 2.5 Flash Image (Nano Banana) came 2025-08-26.
- Date
- Wednesday, 12 March 2025
- Lab
- Google DeepMind
- Kind
- feature
- Access
- research preview
Developer post and API changelog both date it 2025-03-12. Rolled into the Gemini app in April 2025.
Sources
- developers.googleblog.com/en/experiment-with-gemini-20-flash-native-image-generation/
- ai.google.dev/gemini-api/docs/changelog
- blog.google/technology/ai/google-ai-updates-april-2025/
This record was partly confirmed: some claims could not be checked on 6 October 2026. How we check