Free checker · OpenAI and Anthropic
Is your app’s AI being switched off? Check for free
OpenAI and Anthropic are switching off models that thousands of chatbots and automations run on. Paste the error, the model name or a piece of your code: I’ll tell you whether it’s going, when, what to replace it with, what changes in the API, and example code.
Next switch-off: 23 October 2026 — gpt-3.5-turbo, gpt-4, gpt-4-turbo, gpt-4.1-nano, gpt-4o-2024-05-13, o1, o1-pro, o3-mini, o4-mini, gpt-image-1 and their fine-tunes. What to do →
Don’t paste API keys. They aren’t needed, and everything happens in your browser: what you paste is sent nowhere.
- Try:
Check your whole project at once
The same checker, for a whole repository. It runs on your computer with nothing to install: it walks the code and tells you every model being switched off, in which file and line, the date and the official replacement.
In your project’s folder (Node 18 or newer):
npx -y github:ericc2005t/ai-model-sunset- No dependencies: one open-source file (MIT licence).
- Sends nothing anywhere: it reads your files and prints the report.
- The same dates as this page, from the official pages.
And let GitHub tell you
As a GitHub Action: every week it scans the repository and, when a model in your code shuts down within 60 days, opens an issue with the date, the replacement and links to the lines. It closes it by itself when it’s no longer needed. Copy this into .github/workflows/ai-model-sunset.yml:
name: AI model sunset check
on:
schedule:
- cron: '0 8 * * 1' # every Monday
workflow_dispatch:
permissions:
contents: read
issues: write
jobs:
scan:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: ericc2005t/ai-model-sunset@v1
with:
issue-days: 60Switch-off calendar
The dates that break the most apps. The full list is on the official pages.
| Date | Provider | What goes |
|---|---|---|
| OpenAI | gpt-3.5-turbo, gpt-4, gpt-4-turbo, gpt-4.1-nano, gpt-4o-2024-05-13, o1, o1-pro, o3-mini, o4-mini, gpt-image-1 and their fine-tunes | |
| Anthropic | Claude Sonnet 4.5 (claude-sonnet-4-5-20250929) | |
| OpenAI | Agent Builder, the Evals platform and the v1/prompts API | |
| OpenAI | gpt-image-1-mini, gpt-image-1.5 and chatgpt-image-latest | |
| OpenAI | gpt-5, gpt-5-mini, gpt-5-nano and gpt-5-pro (2025 snapshots), o3 and o3-pro | |
| OpenAI | tts-1, tts-1-hd and gpt-4o-mini-tts | |
| OpenAI | gpt-4o, gpt-4o-mini, gpt-4.1 and gpt-4.1-mini (and their fine-tunes) | |
| OpenAI | gpt-realtime, gpt-audio and gpt-4o’s audio and realtime families | |
| OpenAI | whisper-1, gpt-4o-transcribe, gpt-4o-mini-transcribe and /v1/audio/translations | |
| OpenAI | gpt-5.1, gpt-5.3-codex and gpt-5.4-nano |
We’ll email you before your model is switched off
One email 30 days before and another 7 days before the official date, with the replacement and what changes in the code. Free, no account, no marketing.
I’ll fix it within 24 h
If you’d rather not touch it yourself: give me access to the code, I change the model, fix what the new model doesn’t accept and hand it back with how to test it. Fixed price before I start.
Model swap
€79
- One project: a repository, a script or an n8n, Make or Zapier flow
- The new model and the parameters that errored
- Delivered within 24 h of getting access to the code
- I explain what I changed and how to test it
Full migration
€290–490
- When a name change isn’t enough: Assistants API to Responses API, text-to-speech to the Realtime API, Claude Sonnet 4.5 with prefill or tools
- Several services or several calls
- Price and timeline fixed in the quote, by size
- You tell me in the form what your app runs on and what breaks.
- I reply in writing within 24 h with a fixed price and a timeline.
- You give me access to the code: a repository, a zip or the tool the flow lives in. Never your API key: if it has to be tested with it, you set it as an environment variable yourself.
- I hand back the change, with what I touched and how to test it.
No payment on the site: the price reaches you in writing before I start.
Ask for the fix: I reply within 24 h
No calls and nothing to pay here. Tell me what your app runs on and what breaks: I reply in writing within 24 h with a fixed price and a timeline.
By model and by error
- gpt-3.5-turbo shutdown
- gpt-4 and gpt-4-turbo shutdown
- o1, o3-mini and o4-mini shutdown
- gpt-image-1 shutdown
- gpt-4o-mini shutdown
- gpt-4o shutdown
- gpt-4.1 shutdown
- gpt-5 (2025) and o3 shutdown
- whisper-1 shutdown
- tts-1 shutdown
- Claude Sonnet 4.5 retirement
- Assistants API shutdown
- «has been deprecated» error
- «model_not_found» error
- Claude prefill error
Questions
- Is the checker free?
- Yes, with no sign-up. It runs entirely in your browser.
- Do you keep what I paste?
- No. The checker has no server behind it: what you paste is read in your browser and never leaves it. Only what you send in the alerts form or the quote form is kept, if you send them.
- Can I check a whole repository?
- Yes: with the open scanner (
npx -y github:ericc2005t/ai-model-sunset), which runs on your computer with nothing to install and sends nothing anywhere. Also as a GitHub Action, which opens an issue in your repository when a model is about to be switched off. It’s open source, MIT licence. - How do the email alerts work?
- You pick the models you use and we write 30 days and 7 days before each one’s official date, with the replacement and what changes. They’re free, and you unsubscribe by replying «stop» to any alert.
- Are you an OpenAI or Anthropic partner?
- No. Promti has no relationship with OpenAI or Anthropic. The dates and replacements come from their official deprecation pages, linked in every result.
- Can the dates change?
- They can. We check them on the official pages and update them with every announcement. The last check’s date is at the end of this page.
- What if I use Azure, Amazon Bedrock or Google Cloud?
- Those platforms set their own dates, which can differ. Check your platform’s model table.