·2 min read
How this index tracks AI prices, and what it deliberately leaves out
Transparency is part of the product: if a number cannot be reproduced from published sources, it does not belong on the site.
methodologytransparency
Read the article →AI pricing guides and analysis
The data pages on this site answer what does this model cost. These articles answer the questions that sit behind that number: how to turn a per-million-token rate into a monthly budget, why the same model is cheaper on one platform than another, what a support bot costs per conversation, and which pricing terms every rate table quietly leaves out.
Written and reviewed by Mohammad Ghonchegi, editor and maintainer · 10 articles · last updated 2026-10-02. Numbers quoted in the text are generated from the dataset published 2026-10-28 12:57 UTC; see how the index is built.
Articles
10
written and reviewed, not generated
Words
4,414
excluding tables
Latest update
2026-10-02
How this index tracks AI prices, a…
Data behind them
1,557
priced chat models · 32 providers
All articles
Newest first. Every article links back to the data pages it draws on, so you can check the numbers yourself.
·2 min read
How to read an AI pricing page without getting the bill wrong
A pricing page is a table of rates, but the bill is decided by five fields and three footnotes.
methodpricing mechanics
Read the article →·2 min read
Eight ways to cut a token bill, in the order that actually works
Most teams start by switching models. That is usually the fourth-best option available.
cost reductionengineering
Read the article →·2 min read
What changed in AI API pricing: the moves we detected
The changelog is a by-product of how this index is built: each refresh is diffed against the last, so price moves are recorded rather than remembered.
price movementsmarket
Read the article →·2 min read
The same model costs different prices on different platforms
Popular models are resold by many hosts. Grouping the listings reveals a price spread no single provider page will show you.
cross-platformprocurement
Read the article →·2 min read
Why output tokens cost more than input tokens
On nearly every model in this index, generating a token costs more than reading one. That ratio quietly decides which features are affordable.
pricing mechanics
Read the article →·2 min read
The cheapest way to summarise a million documents
When you summarise at scale you are buying input tokens in bulk. Output is almost free; context limits and rate limits are what really constrain you.
summarisationbatch workloads
Read the article →·2 min read
What a customer-support bot actually costs per conversation
Support volume is measured in tickets. Turn the ticket into tokens, and the model choice becomes a visible line in the support budget.
supportcost planning
Read the article →·2 min read
Prompt caching is the biggest lever on your token bill
If your product resends the same system prompt on every call, the cache discount is worth more than any model switch you can make.
cachingcost reduction
Read the article →·3 min read
How to estimate your monthly AI bill before you write the code
Token prices are per million tokens; budgets are per month. Here is the arithmetic that connects them — and the three assumptions that decide the answer.
cost planningmethod
Read the article →How these articles are made
Written and reviewed by a human editor. Prices quoted in the text are generated from the same dataset that drives every page on this site, so they cannot drift away from the tables they support.
- Prose is written and reviewed by a person. Article text is not template output — it is written around the numbers, then checked against them before publishing.
- Numbers are generated, never typed. Every figure inside an article is interpolated from the same dataset that renders the model pages, so an article cannot contradict the table below it. If the dataset changes, the articles change with it.
- Exclusions are stated. Batch pricing, committed-use agreements, regional rate cards and free tiers are outside the normalised rates and labelled as such on every page.
- Corrections go to the pipeline. If a rate is wrong, the fix is applied at the source mapping so it propagates everywhere the number appears — tell us here.
The full recipe is on the methodology page, and the editorial policy sits on the About page.
Full price index
1,557 priced chat models, sortable and searchable.
Cost calculator
Put your own traffic numbers against any model.
All tracked models
One canonical page per model, grouped by provider.