SaaS Platform & AI Model Release Tracking
Model versions, deprecation windows, pricing changes and API-breaking updates across hosted platforms.
Hosted software has no version number you control. The provider updates the service and you are on the new version whether or not you noticed, which inverts the usual relationship between release tracking and action: you are not deciding when to upgrade, you are finding out what changed underneath you. That makes the changelog the primary artefact rather than the download.
AI model releases add a second complication. A model identifier can be retired, aliased to a newer model, repriced, or have its context limits changed — all without the name in your code changing. Deprecation notices for hosted models are effectively deadlines with a fixed date attached, and they are the highest-value events in this category by a wide margin.
What this category tracks
- Model releases and revisions
- New model identifiers, versioned snapshots and capability changes announced by the provider.
- Deprecation and retirement dates
- Scheduled shutdowns of model versions and API endpoints, recorded as dated deadlines.
- API and SDK changes
- Breaking changes, new parameters and client library releases across hosted platform APIs.
- Pricing and quota changes
- Rate changes, tier restructures and quota adjustments that alter the cost of an existing integration.
What matters most here
A retirement date is the only release event with a hard consequence attached. When a hosted model or endpoint is switched off, integrations calling it stop working on a known day. Those dates are usually announced months ahead in a changelog entry that scrolls away within a week, so capturing them as durable, dated records is more useful than capturing the launch announcements that get all the attention.
Where this data comes from
- Official provider changelogs and API release notes
- Published deprecation and model lifecycle schedules
- Client SDK repository release tags
- Public pricing pages and pricing-change announcements
Every indexed entry links to its primary source. See our editorial and sourcing policy for what we verify and what we do not.
Common questions
Why track pricing as a release event?
Because for a hosted service a price change has the same operational shape as a version change: it lands on a date, it is announced in advance, and it requires a decision. Filing it in the same chronological stream as everything else is more useful than putting it somewhere separate.
Do you evaluate or benchmark the models you list?
No. We record that a release happened, when, and what the provider said about it, and we link to the source. Independent evaluation is a different discipline and we do not claim to do it.
Tracked releases in SaaS & AI
The live release list for this category loads below.
Guides covering this area
Building a Release Watchlist You Will Actually Read
Most tracking setups fail within a month for the same three reasons. A practical method for choosing what to watch, how to receive it, and what to deliberately ignore.
LifecycleEnd-of-Life Dates: The Deadlines Nobody Announces
Support windows, LTS branches and deprecation schedules are the only release events where nothing arrives to tell you they happened. How to find them and why they need to be stored, not read.