Hey all, this functionality was removed because some models/providers can cause threads to misbehave if the content differs from what upstream expects. This could invalidate threads and cause other issues.
If providers and models ever get better at handling cases like this, the team could look into restoring the ability to edit LLM responses.
