Hello @CatFly,
Your quota may have been adjusted to 8K TPM based on changes in subscription-specific limits. As an MSDN subscriber, the allocated quota for the GPT-4 series is set at 8K TPM, which may be lower than what you previously had. These adjustments are not random but are part of resource management strategies that vary across different subscription plans.
I attempted to reproduce the issue in my environment and have included a screenshot for reference.
Regarding the unavailability of the O series, it could be due to subscription-based restrictions or regional limitations. The Azure OpenAI Service enforces specific quotas and limits that depend on the subscription type, and certain offer types may have lower maximum quota allocations than others.
please refer this Other offer types.
I hope this helps. And, if you have any further query do let us know.
Thank you!