Home / Add-ons / AI Cost Index Pro
AI Cost Index Pro — plan economics and model capabilities as a dataset
A dated JSON dataset of Plugsky plan prices, intervals and the live model capability matrix, built for procurement comparison and internal budgeting.
Who it is for
Procurement, FinOps and analysts who need a verifiable snapshot rather than a marketing page.
The problem it removes
- Plan pages change and screenshots go stale.
- Comparing an AI plan against a per-token bill needs structured numbers.
- An internal business case needs a dated source it can cite.
What you get
- Every active plan with price, currency and interval
- The same model ladder that serves GET /v1/models
- generated_at timestamp and an explicit methodology field
- No competitor prices unless a dated, verifiable source exists
- Stable keys so a spreadsheet import does not break
How to use it
1. Activate the add-on
Dashboard → Billing → Add-ons → AI Cost Index Pro.
2. Fetch the snapshot
One request returns plans and models together, with generated_at.
3. Pin the snapshot
Store the generated_at value with your comparison so a later reader knows which version you used.
curl -sS "https://plugsky.com/api/addons/cost-index/benchmarks" \
-H "Authorization: Bearer $PLUGSKY_SESSION" | jq '.generated_at, (.plans|length), (.models|length)'
What the dataset contains
The response has two sections and a header that makes the version explicit. Plans come from the billing table that drives checkout; models come from the same ladder that serves the public model list. Neither is hand-maintained in a document that can drift.
| Field | Contents | Use |
|---|---|---|
| generated_at | ISO timestamp | Pin it with any comparison so a later reader knows the version |
| dataset / version | Name and month | Identify the snapshot in a citation |
| methodology | Plain text | States what is measured and what is excluded |
| plans | slug, name, interval, price, currency, limits | Plan economics, including the limits that decide fit |
| models | id, name, tier, context window, features | Capability matrix without a scraped marketing page |
Plan limits are returned as structured metadata rather than prose, so a spreadsheet can compare requests per day, tokens per day, concurrency and requests per minute across plans without parsing sentences.
Using it in a business case
- Fetch once and store the response. A business case should cite a fixed snapshot, not a live endpoint that may change before the document is approved.
- Record
generated_atandversion. Two lines in the appendix are the difference between a verifiable figure and a screenshot. - Join plans to your own volume. The dataset gives price and limits; only you have requests per day and tokens per request. The comparison is the product of the two.
- State the exclusion. The methodology field says competitor prices are not included. Keep that sentence in your appendix rather than implying a broader benchmark than the data supports.
Why competitor prices are excluded
Competitor pricing changes without notice, is often quote-only at the volume that matters, and is frequently published with conditions that a table flattens away. Including it would make the dataset look more useful and be less true. The methodology field states the exclusion explicitly so a reader is not misled by what is missing.
If a comparison against another provider is the actual requirement, the honest method is a dated source per vendor and a note of the date it was checked. That is a research task, not a feed, and it is priced and delivered as one.
Versioning and staleness
The dataset version is a month string; generated_at is exact. Prices change when plans change, and a snapshot older than the current billing configuration should be treated as historical. The response does not claim to be real-time — it claims to be a faithful copy of the billing table at the moment it was generated.
For a procurement document, refresh quarterly and keep the old snapshots. The comparison you made in March is still valid as a record of March, and having both versions is what lets a reviewer see what changed.
A worked comparison
A team sends 1,400 requests a day at 2,800 tokens each — 3.92 million tokens a day, or roughly 118 million a month. The dataset shows Hobby at $8 with a 2 million token daily cap, and Starter at $20 with a 5 million cap. Hobby fails on volume; Starter fits with headroom. The arithmetic is two multiplications and one division, and the appendix can show every input. That is what a dated snapshot is for: the reader can redo the sum without asking you for a spreadsheet.
The same method works in reverse. If a plan looks cheaper per month but its daily cap is below your peak day, the effective price is not the monthly price — it is the cost of the next plan up. Comparing the cap before the price is the order that avoids a migration six weeks after signing.
What it does not include
- No competitor prices. Stated above and in the response itself.
- No contract or discount terms. Those are account-specific and not part of a public dataset.
- No measured latency benchmarks. The models section describes capabilities and context, not throughput.
- No CSV download. The endpoint returns JSON. Converting it is one line in a spreadsheet import or a short script.
Each of those is a plausible product. None of them is this one, and the page says so before the buyer pays.
Licence and redistribution
Internal use is covered by the subscription: dashboards, business cases, procurement documents and board packs. Redistribution — republishing the dataset, reselling it, or embedding it in a product you sell — requires written permission. The distinction is between using a number and becoming a source of the number.
Attribution is not required for internal use, but citing the dataset name and generated_at is recommended, because it makes the figure checkable by the next person.
Who needs it, and who does not
It earns its price when someone has to defend a number: a finance partner building a budget line, a procurement lead comparing two plans, an analyst writing a board paper. It is unnecessary if you only need today's prices for a casual comparison, because the public pricing page already shows them.
The difference is durability. A pricing page is a live claim; this is a dated snapshot with a methodology field, and that is what an appendix needs.
Questions
Does it include competitor prices?
No. Competitor pricing is excluded unless it can be tied to a dated, verifiable source. The methodology field states this explicitly.
How often does it change?
Whenever plan pricing or the model catalogue changes; the response carries generated_at so you can detect a new version.
Can I redistribute it?
Internal use is covered. Redistribution requires written permission.
Is this the same data as /v1/models?
The model section comes from the same ladder. The plan section comes from the billing table that drives checkout, which is not exposed publicly.
What format is it?
JSON over HTTPS. A stable schema with top-level dataset and version fields.
Sources
- Plugsky pricing — checked 21 November 2026
- Plugsky model catalogue — checked 21 November 2026
Related add-ons
Usage Export
Download every request your workspace makes — model, tokens, cost, status and latency — as a CSV with a fixed header. Built for finance reconciliation and warehouse loads.
Spend Guard
A window from 1 to 365 days, grouped by model, by day and by status, so you can find the feature that is eating the budget before the invoice arrives.
Compliance Audit
A scoped technical assessment: what applies to you, what is missing, and the order to fix it. A report and an evidence appendix, not legal advice.
Activate from the dashboard; Stripe handles the subscription and access is granted automatically.
Open billingCompare add-ons