How do we get the ability to use it beyond the 2000 daily tokens? I pay for chatgpt monthly and I have over a dozen active API keys but when i try to use this it runs out. I have no issue paying for the tokens…I just want to use it 
“I really like this API, but it’s just too expensive.”
The out-of-band response is amazing feature! But now it’s difficult to get stable results. Is there some instructions to use it?
I thought the same thing but after using it I think it’s an amazing price. I am running promptswith over 2,000 words along with twilio’s api and its costing me about 10 to 12 cents for every 5 minutes of use. thats at around 75wpm (common conversation speed). i get 30 minutesof talk time per dollar spent using both apis. The average consumer spends 2-3 minutes on the phone with a car dealership, 5-7 with a retail establishment and 8 minutes with tech support. Keep in mind in these scenarios the person they are speaking with has to hunt for answers where as the AI already knows.
You also gotta understand that its not like the normal api, it doesnt send your prompt back and forth along with the replies. it has one initial upload of it and thats it.
I’ve been looking forward to this feature for months I have an early stage project in production leveraging elevenlabs and some open source TTS models with latency being +100ms on average so the realtime voice chat has been in our roadmap, with this I could finally ship it but the current rate limit logic doesn’t allow me to do this, making it completely useless.
I’m currently on Tier 2 which gives me 200 RPM and 40,000 TPM of gpt-4o-realtime-preview and gpt-4o-mini-realtime-preview with this rate I could spent barely 7 - 10 minutes of conversation and then I keep on hitting this:
“Rate limit reached for gpt-4o-mini-realtime in organization org-xxxxx on requests per day (RPD): Limit 200, Used 200, Requested 1. Please try again in 14m24s. Visit https://platform.openai.com/account/rate-limits to learn more.”
Those rate limits are not even feasible for testing and of course neither for production. I know I can pay my way out to higher rate limits but for a kind of acceptable one I’d have to be on Tier 4 basically forcing me to pay $200 in credits to reach it.