How do to ChatGPT-4o Visual narratives?

Looking at ChatGPT-4o Visual narratives demo https://openai.com/index/hello-gpt-4o I am not able to reproduce, what am I missing?

this is the demo

This is what I did to try and reproduce in the chat with ChatGPT-4o

Input

A cartoon mail delivery person with a smile on her face. She is standing facing forward in front of a white background.

Output

Here is the image of a cartoon mail delivery person with a smile on her face, standing in front of a white background. If you need any adjustments or additional details, please let me know!

Input

This is Sally, a mail delivery person: Sally is standing facing the camera with a smile on her face.

output

Memory update (Named the cartoon mail delivery person “Sally.”)
Got it! Sally is now the name of the cartoon mail delivery person. If you need any further modifications or additions, feel free to let me know!

input

This is Sally, a mail delivery person: Sally is standing facing the camera with a smile on her face.
Sally is standing in front of a red door to a house, holding a letter in her hand. We are looking at her from the side.

Output

Here is the image of Sally, the mail delivery person, standing in front of a red door to a house, holding a letter in her hand. If you need any further adjustments or details, please let me know!

The demo looked promising, but I am not able to reproduce - I tried again and it gave me this and the this, what am I missing?


Any ideas :bulb: ?

2 Likes

I think the image-making capability of 4o isn’t rolled out yet.

Here’s what 4o told me today:

OpenAI has recently announced that ChatGPT can now generate images without relying solely on DALL-E. This update is part of a broader enhancement that includes voice and image capabilities integrated directly into ChatGPT. Users can now create images using prompts directly within ChatGPT, leveraging multimodal capabilities from models like GPT-4o and GPT-3.5.

This means that while ChatGPT used to rely on DALL-E for image creation, it now has its own native ability to generate images, making the process more seamless and integrated within the ChatGPT experience. This new functionality is gradually being rolled out to Plus and Enterprise users, with plans to extend access to more users in the near future

For your current requests, I used DALL-E to create the images because the latest features are still being rolled out and may not yet be fully integrated for all users. If you have access to the new features, you might be able to directly generate images using ChatGPT without needing to rely on DALL-E.

How do I know when the new image thing is running on my chatgpt account. And the Sally the mailman can be reproduced?

Is there an update on this? It’s still impossible to generate something similar with 4o even with a subscription.