Standard CCBT indices / Market observations

CCBT Research.

Dated observations. Documented market inputs.
A clear separation between a standard and its evidence.

Understand the core model

Pre-publication · No market observation has been published

Standard CCBT

Market observations

Pre-publication
First observation in preparation

Research begins
with an observable market.

No reference price is published yet. This area is reserved for dated market observations—not fictional inputs from the demonstrations.

What a release will specify
01

A defined market.Qualifying models, measurement scope, and coverage.

02

The two observable inputs.Aggregate market-token usage and standardized model-token prices.

03

A reproducible reference.Usage weights, calculation, dates, sources, and methodology version.

The research record is separate from the standard and all fictional demonstrations.Contribute market evidence

Specify the Market.
Document the Inputs.

The standard defines the calculation. Each research implementation must state which market it represents and how its price and usage inputs are observed.

First specification in development. No qualifying market or pricing convention has been adopted for a published observation in this build.

Qualifying universe
To be specified before publication.
Price and usage sources
To be selected and documented.
Pricing normalization
Input, output, cached, batch, and committed pricing require an explicit convention.
Measurement window
To be stated separately from publication time.
Publication process
Dated snapshots. No live feed is connected.
Coverage and revisions
Each release must describe exclusions and methodology changes.

Published-price estimates and realized transaction expenditure are not interchangeable. The observation will state what is measured and how much of the qualifying market is covered.

General methodology

A Snapshot You Can Inspect.

The research reference is designed around dated, reviewable observations—not a continuously moving price. The source, scope, and conventions belong beside the number.

Scope

A defined market.

Each release identifies the qualifying model universe and any exclusions. A narrower research series must not be presented as coverage of the entire AI market.

Timing

A dated observation.

The observation date, usage window, and publication time are separate fields. An unchanged observation keeps its original dates.

Evidence

Visible assumptions.

Price sources, usage sources, normalization, coverage, and methodology version accompany every published snapshot.

Research is not an executable quote.

A published-price proxy is different from observed transaction expenditure. Any implementation must state what its inputs actually measure and which pricing assumptions it uses.

History begins with observations.

A historical chart appears only when dated research observations exist. The illustrative calculator and fictional time scenario never contribute values to the published series.

There is no live data feed connected in this build.

Review the general methodology

Develop the reference

Better inputs.
A more useful common basis.

Discuss the research