Free · no signup · vendor-independent

The same AI model can cost you 10× more — depending on where you run it.

AI spend scales with usage, so the provider you defaulted to can quietly become a large recurring cost. Before you sign anything, see what each model actually costs across every provider that serves it.

Compare model prices on Verna One →

Why this belongs on the board agenda

For open-weight models, several providers host the identical model and price it independently — so the cost per million tokens can differ by up to 10× for exactly the same output. For closed models, prices move over time and vary by tier and region. The result is simple: unless someone is actively checking, your AI is almost certainly not running on the cheapest sensible option — and no one owns that decision.

Knowing your AI unit economics — cost per task, per user, per outcome — is ordinary financial control. It also protects you from lock-in: when you can see the whole market on one page, switching or negotiating is a data-backed decision, not a leap of faith.

What the free tool shows

Verna One's model price index puts every provider's price for every model on a single page, normalised to USD per million tokens:

• ~1,500 models across 80+ vendors — text, image, audio and video
• The full provider spread per model, so the cheapest and most expensive host are obvious
• Flags on models whose sources disagree on price, and on models being deprecated or sunset
• Free, no signup, runs in your browser, refreshed regularly

How to use it in ten minutes

1. Find your models. Search the model(s) you run today on verna.one/models.
2. Read the spread. Compare the cheapest sensible provider to what you're paying now — the gap is your immediate saving.
3. Check lifecycle. Confirm nothing you depend on is flagged for deprecation, and note the current equivalent.
4. Take it to the table. Turn the numbers into a switch/negotiate decision — that's the governance step.

Frequently asked