I have been working on a project that makes use of the Responses API, specifically tools and files. Unfortunately I am finding it to be unacceptably SLOW. Note the “nano” models are not intelligent enough for my use cases. In any case nano models aren’t even that much faster than mini anyway. And regular models are absurdly slow.
On other project tried (free) Vapi which also allows files and find it to be very fast!
What is the problem? I pay for AI usage but don’t have a pro account, am I being penalized? Why is the API so slow?
You seem unaware that you may have specified an AI model that uses internal reasoning before answering, such as gpt-5.
If you want faster, lower quality seen output, you can employ the Responses API parameter “reasoning”: {“effort”: “none”} to gpt-5.1, or “minimal” to gpt-5 models, and they will think less before a transition to a response.
Well, no, you are not being penalized beacuse you dont have Pro account. API speed generally does not change based on Subscription type. The API si slow mainly because big models and files take time, not because you dont have pro.
Thanks - I removed use of the files tool and replaced with a RAG implementation (which I had been experimenting with anyway) and it is much faster. Other tools seem to be fine, just the files tool has a performance problem.
My files are not complex or large, so I have also experimented with loading them into the prompt and that also has acceptable response times.
I would say OpenAI needs to improve their files tool.