How to Identify the OpenAI Model Used in Your Custom GPTs
Like many, I always wonder with each new OpenAI model release whether my custom GPTs will benefit from it, improve, or degrade in performance based on the model used. Some models may be better suited for coding, while my GPTs might focus on completely different topics.
Keep in mind that you can always ask your GPT directly, and it may respond with its model name. While browsing the OpenAI Dev community, I found answers that I thought would be beneficial to document for better SEO and findability. Here are the steps to follow:
01 — Launch a web browser (e.g., Chrome).
02 — Open a new tab and visit a JSON formatter website.
03 — Open your GPTs in Google Chrome.
04 — Open the Inspect Developer Tools in Chrome.
05 — In Developer Tools, go to the Network tab.
06 — Click on Fetch/XHR to filter requests.
07 — Clear networking logs.
08 — Make sure that the red recording icon is activated.
09 — Use your GPTs to trigger a query.
10 — Locate the Conversation entry under the Name column.
11 — Click the EventStream tab.
12 — Find the Assistant role by using the additional search bar.
13 — Right-click it and select Copy Message.
14 — Paste the content in the JSON Formatter.
15 — Locate the model_slug field to see the model in use.
Et voilà! I have seen various answers to date (02/02/2025), stating that models could be displayed in the GPTs’ description on mobile and/or on the desktop app. However, I realized this was not the case for me, which might be related to feature-flagged functionalities that are not yet available worldwide and/or are currently in beta testing.
One thing that will add value for citizen developers in the coming months and years is the ability to upgrade their GPTs with new OpenAI models to benefit from the latest improvements. At the same time, having the option to stick with a frozen version will help avoid issues such as hallucinations, lack of discipline, and other inconsistencies.
