I’m currently using the Compliance API to programmatically compile and record usage analytics.
Some metrics I would like to compile and present in a dashboard:
- Total use, by user, by GPT (custom/base), by model (GPT-5, GPT-4, variants of each)
- Usage over time (by user per specified time period)
Most of this is currently doable but where I am lacking is the model information.
I want to track which models our users find useful and use most. Ideally the model would be included in the response from the List Conversations API. If the GPT is known (whether it’s a custom GPT or base) the model should also be known.