Offered on models endpoint: "gpt-3.5-turbo-16k"
Chat Completions
“gpt-3.5-turbo-16k” → "model": "gpt-3.5-turbo-0125" returned on call
Responses
“gpt-3.5-turbo-16k” → Error: Model not found gpt-3.5-turbo-16k
“gpt-3.5-turbo-0125” → "gpt-3.5-turbo-0125"
Until 2026-10-23 shutoff, it should be a piece of cake to place the correct pointer so an offered model in the same shape as others on the models endpoint doesn’t have to be gated by chat endpoint. No deprecation notice has been issued explicitly against this model ID.
Note: gpt-3.5-turbo-16k-0613 was shutdown; the replacement ID does not offer the same unlimited output up to end of context window, but rather is 4k max out.