Ability to Verify and Select Sub-Version of GPT Models in Custom GPTs
Description:
I am requesting a feature for OpenAI’s Custom GPTs that allows users to:
- Verify the Sub-Version in Use: Provide a way for developers or users of Custom GPTs to check which specific sub-version of a model (e.g., GPT-4.01) is currently deployed for their custom GPT.
- Ensure the Latest Model: Enable a setting or default behavior that ensures Custom GPTs automatically operate with the most advanced version of a model available (e.g., GPT-4.01 or newer iterations of GPT-4).
- Optional Selection of Sub-Versions: Introduce an option to manually select or lock in a specific sub-version for stability, debugging, or compatibility testing purposes.
Why This is Important:
- Transparency: Knowing the exact sub-version ensures developers are aware of the specific model capabilities their GPT is leveraging.
- Performance: Access to the latest versions guarantees that GPTs are benefitting from the most advanced features, bug fixes, and optimizations.
- Debugging and Compatibility: In some cases, stability may require reverting or locking to a specific sub-version to replicate performance or align with a given use case.
Suggested Implementation:
- UI Integration: Add a field in the Custom GPT management interface to display the sub-version in use.
- API Query: Provide an API endpoint or flag that developers can query to programmatically retrieve the model’s sub-version.
- Default Setting: Enable automatic updates to the most advanced sub-version unless a manual selection is configured.
Thank you for considering this request! This functionality would enhance the user experience, increase transparency, and improve Custom GPT performance for a wide variety of applications.