Independent research, published in full
aiz.to exists because AI vendor marketing moved faster than anyone's ability to verify it. We test, we score, we publish the method.
What we do
aiz.to tracks the AI tools and providers that professional teams actually buy. Every provider on the site is scored on the same fixed suite of tasks, and every score is republished when the suite is re-run.
How the scoring works
Scores are out of 100 across five axes: reasoning, coding, writing, speed and value. Reasoning and coding use verifiable tasks. Writing is measured by edit distance from first draft to publishable. Speed is median time to first useful token under normal load. Value is quality per dollar at realistic volumes.
How often we re-test
A full re-run happens quarterly, with monthly spot-checks on any provider that ships a new flagship model. Model releases move scores quickly, so we date every result rather than presenting a permanent ranking.
Editorial independence
We buy our own subscriptions at retail price. No vendor sees a review before publication, no vendor can change a score, and any page carrying an affiliate link discloses it at the top.
Corrections
If a price, limit or capability on this site is wrong, we correct it and note the change on the page. Accuracy on the boring details is the whole product.
New to the site? The comparison table is the fastest way in, or read the full methodology write-up.