Tokely

THE COST OF BUILDING WITH AI

How we compare AI API prices

The matching rules, workload formula, source dates and exclusions behind Tokely's price tables. Understand reference estimates, recorded savings and media comparisons.

Try in Playground Calculate your cost

Published by Tokely · Reviewed

THE SHORT ANSWER

We match exact model identifiers, compare the same token workload and disclose each source date. Missing references remain unavailable; negative differences remain visible. A current price estimate is not historical customer savings, and different media variants are not treated as equivalent.

Exact identity, explicit provenance

Tokely retail rates come from the same versioned catalog used by its model pages. OpenRouter references are matched through existing reviewed model mappings against the public Models API. We do not infer a match by similar names, substitute another model revision or treat missing prices as zero.

One formula, a disclosed workload

The default text table prices 1,000,000 uncached input tokens and 250,000 output tokens. Difference equals (reference cost minus Tokely cost) divided by reference cost. Positive values mean the Tokely estimate is lower; negative values mean it is higher. The calculator lets you change the mix and cache-read subset.

Scope and exclusions

Token comparisons exclude checkout/platform fees, taxes, special tool and media surcharges, cache-write charges, long-context tiers and negotiated contracts. Reference rates do not certify identical quality, region, privacy or latency. Media tables preserve each variant's unit and parameters and do not compute a cross-variant savings percentage.

Recorded savings and corrections

The dashboard's recorded savings uses an official baseline frozen at request time when comparable units are available. Current-rate calculators are estimates. Historical requests without a valid baseline are not retroactively verified. Missing or incomparable usage lowers coverage rather than becoming an invented savings number.

Review and ownership

Tokely publishes these comparisons and benefits when a reader becomes a customer. Data updates must preserve source URLs, dates and matching rules, and must show price increases as well as decreases. Report a mismatch to [email protected] with the model ID and source URL so it can be checked.

Reproduce a comparison

uncached_input = total_input - cached_input
cost_usd = (uncached_input * input_rate
          + cached_input * cached_input_rate
          + output_tokens * output_rate) / 1_000_000
difference_pct = (reference_usd - tokely_usd) / reference_usd * 100

Use the JSON dataset with the dates attached to each source. OpenRouter rates are collected from its public Models API; its plan fees are separate.

Questions before you switch

Why not claim the lowest price everywhere?

The dataset does not cover every supplier, negotiated contract, tier or workload. A defensible comparison describes the exact model, inputs, units, dates and exclusions.

Does a positive difference prove actual savings?

No. It is an estimate for the stated workload. Actual savings require comparable historical usage, settled charges and a valid request-time baseline.

Keep exploring

AI API pricing. Know what you pay. ↗

Published Tokely rates for text, images, video and audio. Pay from one USD balance, compare exact models and estimate your workload before you switch.

A cost-focused OpenRouter alternative ↗

Compare exact-model rates, supported endpoints and migration requirements before moving to Tokely. One balance for supported text, image, video and audio models.

AI API price index ↗

An exportable first-party dataset of enabled Tokely models, exact-model OpenRouter token references and supported media variants. Sources, units and dates are included.

OpenRouter migration checklist · Billing documentation

START WITH YOUR OWN WORKLOAD

Know the price.
Then make the call.

Check the model, compare the rate, and connect with a Tokely key.

Read the quickstart ↗Explore all models →