Skip to content

Models & products

Same GPT-4o name, different shutdown dates: which version does your app use?

3-line summary

  1. October 23, 2026 is the scheduled shutdown date for a group of older OpenAI API models.
  2. GPT-4o versions have different deadlines, making the exact model ID important.
  3. Official model lookup documentation exposes an optional shutdown_date field.
PracticalDevelopers
A magnifying glass examines two model blocks with different calendar markers on white cards against a blue background
Illustration: AI Ping

OpenAI’s deprecation schedule assigns October 23, 2026 to API models including gpt-4o-2024-05-13 and gpt-4-0613. The timetable was announced on April 22. It concerns specific older models, rather than a simultaneous end for every GPT-4 variant.

The same schedule puts gpt-4o-2024-08-06 and gpt-4o-2024-11-20 on January 14, 2027. The GPT-4o model page describes snapshots as names that pin a particular version. A broad product name and the version an application specifies are different things.

Model list and retrieval references define shutdown_date as an optional date or null. Null means no shutdown date has been announced, not perpetual availability. The deprecation page also distinguishes advance notice from actual loss of access. This report checks the published schedule; it did not test shutdown times or account availability through API calls.

Why it matters

AI Ping sees exact model versions as part of routine operations. Replacing a model should include checks of the application’s inputs, tools and output formats alongside the name change.

Who it's for

Developers
Compare configured model IDs with those used in requests, then review the relevant deadlines and replacement candidates.

Sources

  • How-to guideAbout 14 min reading

    Your first LLM API call: from key to response

    Using the Claude API as the example: keep your API key safe, send a first request from Python, and read the answer, the stop reason and the token usage.

    BeginnerDevelopers
  • How-to guideAbout 15 min reading

    Get structured (JSON) output from an LLM

    Instead of asking for JSON and hoping, define the shape as a schema and get responses your code can use directly. Worked through with a customer-message classifier.

    Intermediate · practicalDevelopers