How to replicate Sora's 'remix' feature with multi-turn image generation via API?

API
,
1

Moved to new thread:

3

Also curious about this! I’m finding that image editing in Sora is significantly better than gpt-image-1, and I can’t recreate the same consistency with the API at all.

Does Sora use gpt-image-1 or is there occasional testing of newer models that are not available via API yet?

1 Like
4

This topic was automatically closed after 4 hours. New replies are no longer allowed.