OpenAI API separates slow_down 429s from server_is_overloaded 503s
UpdateVerifiedAdded Sep 22, 2026
Traffic ramping too quickly now returns 429 slow_down, while temporary model overload returns 503 server_is_overloaded. Honor Retry-After when present, otherwise use exponential backoff.
Topics: Developer tools
Summaries of vendors' own notes. Product names and logos belong to their owners; logos via logo.dev.
More OpenAI API releases
Every OpenAI API release| Date | Release | Type |
|---|---|---|
| Sep 3 | OpenAI releases GPT-6 Astra for complex end-to-end work Update | Update |
| Sep 3 | OpenAI Responses API adds async tool calling and mid-turn steering for GPT-6 Astra Update | Update |
| Sep 1 | OpenAI API supports IPv6 connections to api.openai.com Update | Update |
| Aug 29 | OpenAI API mTLS and X.509 workload identity federation are GA GA | GA |
| Sep 8 | OpenAI Prompt Cache Diagnostics is generally available in the Responses API GA | GA |
| Sep 8 | OpenAI releases GPT Image 2.5 Sunburst and GPT Image 2.5 Flare Update | Update |
| Sep 8 | OpenAI GPT-Rosalind is GA for approved life sciences research GA | GA |
| Aug 26 | OpenAI deprecates whisper-1 and gpt-4o transcribe models, shutdown February 26, 2027 Deprecation | Deprecation |
