[Screenshot 2025-08-08 at 14.32.47] I’m using Agents SDK with Responses API. I was testing gpt-5 before switch and noticed that it takes around 1 minute for even a basic query. The Traces shows that: Step Model Input Tokens Output Tokens Duration Triage Agent 1st call gpt-5-nano ~3k ~2k ~…

This was super helpful - thank you. Turned out my blank responses were because the reasoning tokens were hitting 2048 which was the apparent default output allowed. Once I bumped max_completion_tokens up to 5000 GPT-5 can’t stop talking. :slight_smile:

I had it set to minimal and even then it was producing slow, unreliable responses.

5 is also very slow for me and breaks tool calling.

Still slow for me today. 4o and 4.1 responds near instantly, where as 5 and 5 mini take around 30-60seconds to respond. Im just using basic set up from vercel docs. import { openai } from '@ai-sdk/openai'; import { convertToModelMessages, streamText, UIMessage } from 'ai'; import { NextRequest } …

Why was this flagged. I love ChatGPT and I really really want this to work but using it everyday it has become unbearable. It goes back and forth. Sometimes it somewhat fast and by fast I mean at most 10 seconds for a response which is already not good to waiting for 5 minutes for a response. Do you…

[image] lairny: Why was this flagged Look up top: [image] Everything you see makes talking about the consumer ChatGPT product off-topic. ChatGPT not a thing that API-using developers can help you improve, except to tell you to now pick “mini” over there in the new selector, for faster star…

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

Good to hear. I thought it was my code :slight_smile: ”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 “verifi…

on top of that quality is far worse than old models

GPT-5 + Responses API is extremely slow

API

ram9 September 4, 2025, 12:52am 34

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

Topic		Replies	Views
GPT-5 is very slow compared to 4.1 (Responses API) API gpt-5 , reasoning , gpt-41 , responses-api	65	39408	June 1, 2026
GPT-3.5 Turbo API response is slow API	20	12903	November 11, 2023
GPT-3.5 API is very slow. Any fix? API	31	10447	October 12, 2023
GPT-4.1 models are very slow due to API response. API	5	1113	September 8, 2025
Extremely slow API responses and hanging API	12	989	September 15, 2025

GPT-5 + Responses API is extremely slow

Related topics