Setting reasoning effort to low doesn’t make any big difference in performance. It still very slow and its actually reasoning more compared to the GPT4-o despite that the reasoning effort parameter is set to low…
You can set it to “minimal”, it’s a bit faster, but then results for me are worse than with 4o or 4.1.
I’ll stick to 4.1 and just wait to see what happens
1 Like
Good to hear. I thought it was my code 
”Tell me a joke”, responses API, GPT5 nano, json requested, took 4 to 5 sec. And the joke was not even funny.
My take: OpenAI does not want developers to use their API:
- Restrictive rate limits,
- changes on API every two minutes,
- forced “verification” (what for?),
- and slow.
1 Like
on top of that quality is far worse than old models
1 Like
p50 for a structured JSON response is x5-x10 compared to 4o when setting reasoning to minimal. This is unusable.
the joke was not even funny.
this was personal
1 Like
Facing similar Issue here, Glad I came past your comment.
1 Like
Same issue here. My requests range from 12k to 120k input tokens: I’m getting over 30 minutes for a response consistently.
1 Like
same issue. Responses API is extremely slow regardless of model choices or previous_response_id. Refer here as well Stateful Responses API Much Slower Than Chat Completions - #5 by stevecoffey
Iv been using Chat GPT5 for a month now and it’s been fab , until Sunday 7th aug 25 pm and then it went to rat poop ! Waiting ages for response and leaving me right in the middle of something …So pee’d and disappointed…
No its pretty slow crashes in deeper chats mobile ok with fresh chat and taking pics of my shopping list but my main chats on work and improving my excel databses is harrrd work
I faced similar issue. I’m looking for a recent model to replace gpt-4o agents. I
then decided to use gpt-5-chat (in Azure OpenAI) which support function calling while gpt-5-main (in OpenAI) doesn’t. And it performed quite well and with the same speed as 4o. It though tends to avoid calling tools
Same issue here.
i have a langgraph agent workflow, which takes about 5-7 sec in 4o / 4.1
takes 22 sec in gpt 5 mini. (340 % slower)
I can also confirm GPT-5 is super slow (using it from Eastern Europe through Cloudflare AI Gateway).
I can confirm that switching to gpt-5-chat-latest model is SIGNIFICANTLY faster.
GPT-5 is too slow as is, Claude Sonnet 4.5 is MUCH faster BTW
1 Like