Hi team,

I am planning our production workload using AI Gateway on the paid tier and would like to confirm the current capacity, rate limits, and concurrency limits for the following two models:

1. xAI Grok STT
2. Microsoft MAI-Transcribe-2 streaming

Could you please clarify:

• The current Requests Per Minute (RPM) limits for each model
• The maximum number of concurrent streaming sessions or active connections supported
• The process for requesting custom rate limit increases or capacity reservations for higher-volume workloads

Any recommendations or best practices for handling audio streaming concurrency through AI Gateway would also be very helpful.

Thank you!