I wanted to bring to your attention a significant issue many have likely been encountering without knowing. OpenAI usage stats are showing erroneous API responses for GPT-4-Turbo models across both the API and Playground environments for the following models:
- gpt-4-turbo
- gpt-4-turbo-2024-04-09
- gpt-4-turbo-preview
Usage statistics suggest that these models have been used multiple times from my account, yet it appears no actual processing has occurred from these models. The knowledge cutoff continues to be reported as September 2021, and knowledge tests for significant world events after that time and before December 2023 seem to confirm this. (I have sufficient funds in the account also)
Submitted a ticket to OpenAI but the estimated response time is 3 days. Hoping more people can verify so this issue can be escalated quickly.
cheers
- Tyler