How We Review AI Models and Check Prices

AIVideos.eu is run by Richard Byrne, who has 25 years of photo editing and creative design experience, has worked freelance since 2015 and has produced work with AI image and video tools since 2021. He founded AIVideos.eu in 2025. More on the about page.

We also operate the AIVideos Studio, where the image and video models we write about can be run from one credit balance. That matters for this page twice over: it is how we can render the models we review, and it is a commercial interest you should know about when we compare the Studio with other services.

Every provider price we publish is read from that provider's own pricing page and stored in a single cost table with the date it was last checked. The most recent full check was 2026-09-26. The Render Ledger and our AI video cost per second research are generated from that table when the site is built, so when a price is corrected there, those pages change with it.

Where a provider publishes a token formula instead of a per-second rate, we apply their formula and mark the result as an estimate. In our cost research, launch discounts are left out and the standard rate is shown.

The same table sets the floor for our own credit prices: the Studio never charges less for a render than the provider charges us for it.

Our newer reviews of models we run carry a “Key facts” box: clip lengths, resolutions, aspect ratios, input options and the Studio credit price. That box is read from the Studio's live configuration when the page is built, not written into the article, so it cannot drift from what the Studio actually offers. When we retire a model, its box disappears rather than going stale.

A provider's documentation tells you what a model is meant to do. A render tells you what it did. Our standard is to keep the two apart: when we describe how a model behaves, we say whether it comes from our own renders or from the provider's documentation, and we do not present a provider's claim as a test result.

We do not publish star ratings or numeric scores for models. A model that is right for one shot is often wrong for the next, so we write about what each one is good and bad at, and at what cost.

AI models and prices change quickly. When we correct a fact or re-check an article, its “updated” date changes; when nothing has been reviewed, the date stays as it was. Some of our older articles were written before this process existed, and we are re-checking them against the same sources. Where we find an error, we fix it.

If you spot something wrong, a price, a specification or a claim, please tell us and we will check it.

We earn money in three ways: Studio credits, professional AI video production, and affiliate commissions from some of the tools we write about. Any article that contains affiliate links says so in a disclosure at the top.