About
AI Model Overview tracks what happens to LLM models after you choose one: when they get deprecated, when they get switched off, and what changes in between.
Why this exists
Picking a model is well served. There are pricing tables, benchmark leaderboards and comparison posts for every combination you could want. What is missing is everything that happens afterwards — and that is the part with a deadline attached.
Providers do announce deprecations and shutdowns. They announce them in changelogs, migration guides, doc footnotes and blog posts, each on their own schedule and in their own words. If you use four providers, that is four places to watch and no shared calendar. The usual result is that a shutdown is discovered rather than planned.
So we built the boring thing: one record per model, with its status, its announced dates, its price, and a link to the provider page it came from. Today that covers 354 models across 20 providers, 47 of which are deprecated and 71 already shut down.
How we work
Everything comes from published provider documentation, is stored with its source and a verification date, and is validated before it is published. Where a provider has not announced something, we leave the field empty instead of guessing. The methodology page spells out the whole pipeline, including what we deliberately do not do and where the dataset currently falls short.
The data is free to read, on the site, in the overview app and through MCP at https://mcp.aimodeloverview.com. No provider pays to be listed or ranked, and there is no advertising —
so nothing has to be shaded to keep anyone happy.
Who is behind it
AI Model Overview is built and run by Digital Footprint, a small independent studio in the Netherlands. It is not backed by, affiliated with or endorsed by any of the model providers listed on this site. The first public version went live in July 2026 and is actively developed.
Get in touch
Three things we always want to hear about, in this order:
- A wrong lifecycle date. Send the model and the provider page that contradicts us. This jumps the queue over everything else.
- A provider we do not cover. Tell us which one and roughly how you use it — coverage is prioritised by actual demand.
- What you would want an alert to do. Alerts are being designed now, and the people who have been burned by a shutdown have the sharpest opinions about them.
Reach us at hello@aimodeloverview.com.