Tokens are returned from the chat completions API in big, choppy batches. They used to be smoothly returned one by one. See this example import time from openai import OpenAI client = OpenAI() response = client.chat.completions.create( model="gpt-4o", messages=[{"role": "user", "content"…

Thank you so much for reporting the issue! We’ve done some investigation, and this was caused by a configuration change on our end, which we have now reverted. We also recommend using our official library for streaming as robust client code can help protect against this. Sorry for the trouble, and…

Streaming events returned bunched up

API Bugs

PaulBellow July 17, 2024, 9:08pm 3

Might be related to this…

Do you mind if I merge the threads?

Topic		Replies	Views
4o and 4 API output has typo/missing words Bugs gpt-4	55	837	July 19, 2024
Malformed streaming answers from GPT-4 completions API lately API	11	2331	November 13, 2023
Chatgpt api (openai-node v4.26.0) stream issue with gpt-4 models Bugs	18	1449	February 15, 2024
BUG: Streaming packets changed Bugs chatgpt , api	23	2984	November 1, 2023
Incomplete stream chunks for completions API API api , completions	8	3248	September 25, 2023

Streaming events returned bunched up

Related topics