I can’t find anything official about the concurrent session limit or the RPS limit for the Live API models (gemini-flash-live 3 preview and Gemini 2.0 Flash Live native audio). Gemini and ChatGPT claim “3 in AI Studio, ~1000 in Vertex,” but neither can give me an official source.
I run telephony infra that handles 10k+ concurrent calls and I’m connecting them to the Live API, so I need real numbers before scaling.
I did run a small stress test on AI Studio (RPS 2, 10 concurrent sessions) and got no 429 errors — but honestly, why should I have to discover these limits by burning my own quota? This should just be documented somewhere.
Could someone please point me to:
- The official concurrent session limit (AI Studio vs Vertex)?
- The official RPS limit?
- Whether these are hard caps or quotas that can be raised?
Thanks!
Ok finally found something:
«Bidi generate content concurrent requests per project per region per base model» = 5,000
(for gemini-live-2.5-flash and gemini-live-2.5-flash-native-audio). No info for gemini-3-flash-live
Hi Jason! We currently do not guarantee support for concurrent sessions because any limitations we have are on the Tokens per minute basis.
You can request an increase by filling out this form and letting us know how many tokens per minute and per day you plan or need to consume.
Please let me know if you need any further support. Thank you so much!
Facing a similar issue with gemini-3.1-flash-image. my app triggers up to 10-15 images in a window of ~1 minute. I’m on tier 1. I’m not able to get all image generation calls successful in the google Agent platform. but it goes successful in the ai studio. I wanted to move fully to agent platform but this concurrency limit is stopping me. Is there anyway we can increase it in the agent platform? And I don’t see the form you mentioned to increase the concurrency. Please help