Sudden jump in price for OpenAI
# support
j
As of yesterday, we've seen a significant spike in price on our dashboard for all the OpenAI models (~3x) on specific assistants of ours. We have made no changes on our end. Some of our prompts are relatively large, and I notice this price increase applies to our large prompt assistants and not our small prompt assistants. That being said, we haven't made any changes and we're not sure what has caused the price to nearly triple over the course of a couple days. What could explain this?
v
Hi @josh, I'm an AI assistant for Vapi. I'll analyze available sources and get back to you shortly.
----------------------------- **While I generate an answer, here are some sources that might be helpful**: * [Web Calling]() * [Dashboard]() * [Estimating Costs]() * [Outbound Calling]() -----------------------------
### Answer The sudden increase in pricing for using OpenAI models in your Vapi assistants could be due to several factors, including changes in the underlying costs of the model providers or adjustments in Vapi's pricing structure. It's essential to ensure that your token usage per prompt is optimized, as longer prompts can significantly increase costs due to higher computational requirements. Additionally, Vapi uses provider keys you set for communication with vendors like OpenAI, and changes in the API or pricing by OpenAI could also reflect in your Vapi usage costs [(1)]() . If you're seeing discrepancies or unexpected increases, it's important to review the detailed cost breakdown available in your Vapi dashboard to understand which part of your voice pipeline (STT, LLM, TTS) is contributing most to the cost. From there, you can make more informed decisions about potential optimizations or need for changes in your assistant configurations. For a definitive explanation and to address this specific issue, it would be beneficial to directly consult the detailed billing and usage reports on your Vapi dashboard and compare them with the pricing models provided by OpenAI and other vendors. If this does not clarify the reasons for the price changes, contacting Vapi support directly might be necessary to get specific insights into the pricing adjustments or any potential billing errors.
j
Also we are seeing the same for 11labs - a jump from roughly .04 to now .072. Has there been an increase there as well?
s
Can you share your call_id?
j
715db69b-258e-49f0-b990-032a46746f94
s
Okay, so basically the amount is not increased or anything, as per, you know, those dashboard things are the approximation, correct? So, basically, what happens is whenever a call occurs, your prompt is sent and then your conversation, and they keep getting added to your API request so that the LLM model has previous knowledge about the conversation. So, due to that, you will be charged more compared to someone who has a small prompt. So, it is just to give you a better approximation idea about how much you might have to pay, but it is not the actual cost. It is just increased to provide you a better understanding about it. You will still be charged at cost, which means the exact same price you pay for OpenAI requests.
p
also seeing the same thing, is it because of 11labs ios charging more now ?
s
Please read the message once
p
got it so it's the estimation that is more accurate? Im not having a problem with openai , just the 11labs voice is now showing as more expensive .
s
Yup.
That’s not possible because we didn’t changed any pricing model or anything we just made our estimation model better
3 Views