gpt-realtime
gpt-realtime deprecation date: Jan 20, 2027, replacement gpt-realtime-2.1.
Countdown
64 days
Jan 20, 2027
Price per 1M tokens
$4 in · $16 out
Standard tier, list price
Price comparison
Full migration notes: gpt-realtime vs gpt-realtime-2.1
Timeline
Deprecation announced
Jul 20, 2026
Shutdown
Jan 20, 2027
Source: https://developers.openai.com/api/docs/deprecations · checked Nov 16, 2026
Change history
Related OpenAI models
- RetiredSep 13, 2024
- RetiredSep 13, 2024
- RetiredSep 13, 2024
- RetiredDec 6, 2024
- RetiredDec 6, 2024
- RetiredJun 6, 2025
- RetiredJun 6, 2025
- RetiredJun 6, 2025
- RetiredJul 14, 2025→ gpt-4.1
- RetiredOct 10, 2025
FAQ
When does gpt-realtime shut down?
OpenAI lists Jan 20, 2027 as the shutdown date for gpt-realtime (64 days from today). After that date, API calls to it fail.
What should I replace gpt-realtime with?
OpenAI recommends gpt-realtime-2.1. At list prices that is a +29% change in blended cost per token.
How much does gpt-realtime cost per 1M tokens?
$4 per 1M input tokens and $16 per 1M output tokens at the standard tier, from OpenAI's pricing page.
How do I find every place my code calls gpt-realtime?
Paste your code or a public GitHub repo URL into ModelRot's free scanner. It lists every file and line that uses gpt-realtime or any other retiring model, with the replacement and cost change.