Have priority turned on in project.
Seeing this in usage..
But in completion, the info (logs) shows nothing useful or anything about it..?
Are they being processed on priority or not? How to know?
3 Likes
Check the service_tier field in the chat.completions response body. This field reflects the tier that actually served the request and may differ from what you asked for. So a request only definitively used Priority processing if the returned service_tier is priority, which it looks like it did in this case.
https://developers.openai.com/api/docs/guides/priority-processing/
https://developers.openai.com/api/reference/resources/chat/subresources/completions/methods/create/
1 Like
Unfortunately I don’t have access to that since I’m plugging in via another service.
Curious, since i was previously looking for 5.4-nano + priority, and reason it didn’t exist was since it wasn’t on the pricing - but this looks like it could be being served by that?
1 Like
I recently came across the same issue regarding 5.4-mini and -nano but these models are not served via priority processing.
You can check this for every model on the pricing page when selecting priority:
In your case with the 4.1 models I would trust the usage data from the platform.
Hope this helps!
2 Likes
gpt-4.1-mini does exist as a “priority” model, and you can also get a zero bill if enrolled in complementary daily tokens. The notable exception without priority service tier is gpt-5.4-nano.
Lots of zero billings if you click “other” on the legend of Usage.
The peculiar thing I note: despite a $0 total for “priority | gpt-4_1-mini”, it appears with its own graph, indicating it was called (models not used will not be shown), but with no data. This is a fairly useless UI, and what can be considered the confusing issue, bringing about this forum posting.
The fault is:
- usage → spend categories
- see tons of models appearing despite no dollar bill
- the graph is there if you hover, to see also a daily $0
