Description
Gemini just released image outputs natively in their API:
https://developers.googleblog.com/en/experiment-with-gemini-20-flash-native-image-generation/
We should support this in a provider/model-agnostic interface (since OpenAI and others have hinted at this and I generally expect this to be more common across various models and providers)
Description
Gemini just released image outputs natively in their API:
https://developers.googleblog.com/en/experiment-with-gemini-20-flash-native-image-generation/
We should support this in a provider/model-agnostic interface (since OpenAI and others have hinted at this and I generally expect this to be more common across various models and providers)