OpenAI (ChatGPT, DALL·E, Whisper) – Generate Image (gpt-image-1)
Overview
We have recently launched a new action under the OpenAI (ChatGPT, DALL·E, Whisper) integration:

This action enables users to generate images using the latest OpenAI image generation model (gpt-image-1).
Reported Issue
A user reported encountering a 403 Forbidden error when attempting to use the gpt-image-1 model, as shown in the error log:
- Error Message: Error from upstream: Forbidden
- Error Status: 403
- Error Code: UPSTREAM_ERROR
- Error Source: cf-wk-receiveactionfiledata
This clearly indicates that the test request fails at the authorization level.
Root Cause
Upon investigation and user feedback, the root cause was identified as follows:
The user’s OpenAI account was not verified to access protected models like gpt-image-1.
Access to certain models, including gpt-image-1, requires the OpenAI account to be linked to a verified organization via eKYC verification. This verification ensures responsible use of advanced capabilities such as image generation.
Resolution
Once the user completed the organization verification process via Persona (OpenAI’s identity verification service), access was granted, and the action started working as expected.