Simple to read.
Explicit about the evidence.
The Intelligence Hours Index (IHI) measures the gross minimum-wage working time needed to pay the official USD monthly list price of an AI subscription. It translates a dollar price into gross earnings time at the applicable wage floor. Actual purchasing power also depends on consumer taxes, take-home pay and whether the wage is paid. It does not rank AI model capability.
The proposal is a new Big Mac Index for the AI age: a familiar purchase that makes an international economic relationship visible. The question is what local labor can buy at a dollar price. Wages and exchange rates jointly determine the answer. Productivity gains matter to this measure when they reach the applicable wage floor or exchange rate. Read the motivation and primary sources.
Intelligence Hours: Dollar-Priced AI and the Purchasing Power of Labor
The paper studies 37 countries over 2023–2025, with a 2026 extension to 12 economies. It connects subscription prices to statutory minimum wages, includes a Kaitz-ratio wage-denominator scenario, decomposes exchange-rate and wage movements, and examines local pricing within the same plan. The historical common-price series uses a pre-tax USD 20 monthly reference. This public v5 edition audits dated wage schedules and their units, adds consumer-tax scenarios, compares local prices on a consistent tax basis, and revises statistical inference for repeated countries and shared currencies. The country-level Claude usage check moves to supplemental detail. This edition is distinct from the journal-submission manuscript.
Research sample and findings. The study uses a 111-row country-year panel for 2023–2025, with calendar-year wage references and period-average exchange rates. Most references use selected adult basic pay, with supplementary payments kept separate. Canada's and Japan's historical values retain employment-weighted regional benchmarks; Belgium's guaranteed average monthly income has a distinct remuneration scope. These exceptions do not establish a uniform national hourly floor. For 2025, the paper reports 36 nonmissing market-rate observations, ranging from 1.14 to 16.20 hours at the USD 20 reference price. After the dated wage and currency-unit audit, the income elasticity is −1.03 (HC3 95% confidence interval [−1.23, −0.83]). The pooled estimate is −1.06, with a country-clustered interval of [−1.22, −0.89] across 110 usable observations. Both intervals include −1 without establishing exact inverse proportionality. Country clustering handles repeated observations within a country; shared currency or regional shocks can still create dependence across countries. The local-price pilot distinguishes listed prices, tax treatment and the timing of the US reference price; these are not independently verified local checkout observations. Replication data and code are published in Zenodo, with the historical panel, frozen source inputs, analysis scripts and exploratory extensions.
Versioned evidence. Zenodo 1.2.0 and the v5 supplement include dated country-source wage corrections and their conversion policies, country-level official tax and net-pay audits, conditional payroll profiles, revised statistical analyses, the qualified January 2026 snapshot, and exploratory AEI inputs and code. The original Zenodo 1.0.0 deposit preserves the earlier data and code; the earlier Zenodo 1.1.0 revision preserves the 2 October tax and statistical edition. The country-source corrections belong to version 1.2.0. Read the concise external-criterion check.
Ahn, K. (2026). Intelligence Hours: Dollar-priced AI and the purchasing power of labor (Public version 5, country-source correction, 3 October 2026; argument revised 5 October 2026; title revised 6 October 2026) [Research paper]. INLEVEL9.
Data and code citation. Ahn, K. (2026). Intelligence Hours: replication data and code for the public v5 country-source correction (1.2.0) [Data set]. Zenodo. DOI: 10.5281/zenodo.23105789
Title Page Information
Kwangseob Ahn, Sejong University, Seoul, Republic of Korea. ORCID: 0009-0000-5299-1457. Correspondence: oswarld@sju.ac.kr.
Declaration of interest: none. Funding: no specific grant from public, commercial or not-for-profit funding agencies.
Data availability: public v5 country-source correction, version 1.2.0, https://doi.org/10.5281/zenodo.23105789. The DOI identifies data and code, not an accepted journal article. The original release manuscript, separate supplementary analysis, frozen inputs, source captures, scripts and SHA-256 manifest are included in the replication release at https://www.intelligence-hours.org/research/validity/v5-revision-replication.zip. Prior versions remain linked in the repository's version history.
CRediT author statement: Kwangseob Ahn: Conceptualization, Methodology, Data curation, Formal analysis, Writing – original draft, Writing – review and editing.
Public version 5, country-source correction, 3 October 2026; argument revised 5 October 2026; title revised 6 October 2026. Word count is reported in the reading edition's formatting verification.
Version Note
The 3 October correction reconciles the 2023–2025 wage references with dated national sources under one disclosed averaging policy. It preserves the original provider inputs and separates ordinary base pay, mandatory extras, regional aggregates and directly notified hourly tariffs. Bulgaria's EUR/BGN mismatch is corrected before wage harmonization; the 2025 bridge in Supplement F distinguishes currency alignment from subsequent rate and scope choices. Japan's January cutoff and Poland's employment category are corrected in the separate snapshot; China's and India's unharmonized national wage ranges are withdrawn. Unmatched Kaitz products are conditional source-ratio scenarios, not measured median earnings. The 5 October reading revision brings the author's three-part argument into the abstract, introduction, framework, discussion and conclusion. It retains the corrected estimates, tables and figures. The DOI continues to identify the frozen 1.2.0 data/code release; this revised reading text is not a replacement of that deposited archive. The 6 October update changes only the paper's title and publication metadata. The 37-country, 2023–2025 study, body text, estimates and figures remain unchanged from the author's designated final manuscript.
The paper's 1 January 2026 wage snapshot.
The extension retains 12 economies and separates each wage's worker category, jurisdiction, published unit and effective date. Its dollar-price scenarios use the frozen 2025 annual exchange rates. A monthly wage divided by 173⅓ hours uses an analytical 40-hour week; it is not a government-notified hourly tariff. The same hours convention expresses subscription costs as a share of normalized monthly minimum-wage income. Missing national wage values remain unavailable while their separate GNI comparisons can still be calculated.
- Japan: the general prefectural floors in force on 1 January range from ¥951 to ¥1,226 per hour. Akita's new ¥1,031 rate takes effect on 31 March 2026. The later fiscal-year weighted average is not a verified point for the snapshot date, so the national point and its Kaitz-rescaled comparison are withheld.
- Poland: ordinary employment uses the PLN 4,806 monthly minimum, normalized to 173⅓ hours. The PLN 31.40 hourly minimum for specified civil contracts is a separate worker-category sensitivity.
- China and India: national wage ranges and points are withheld. Xinjiang's CNY 17.5 and Beijing's CNY 27.7 regional part-time floors have different contribution definitions. Delhi's ₹710 daily unskilled floor is a dated regional illustration from April 2025; its exact continuation at the January cutoff is not independently established.
- Nigeria: official statements corroborate NGN 70,000 per month. The hourly equivalent assumes 173⅓ hours and a legally covered employee; consolidated statutory coverage and exemptions remain unresolved.
These are research wage comparisons against a common price. They do not establish observed local checkouts, take-home wages or wages received by every worker. Read the dated official sources and scope limits.
How the index is calculated.
Intelligence Hours = monthly USD price × local currency per USD ÷ local hourly wage
The verified index uses this basis: US list price · before taxes · gross minimum wage. Reference exchange rates change over time; the USD pricing convention stays fixed. “Fixed USD” does not mean freezing a provider’s price forever.
The USD basket is a common reference price chosen for comparison. It does not mean that every customer pays the same amount worldwide. Providers can offer regional prices, different tiers, taxes and payment conditions. The index asks how much local work buys the same published US price, not what a visitor will see at checkout.
An index measured in hours.
An IHI of 2.5 means two hours and thirty minutes of work for one monthly subscription. The score retains this direct meaning; there is no arbitrary 100-point baseline. Rank 1 has the fewest required hours among verified capital observations. Ranks use unrounded values, and exactly equal values share a rank. Changes in coverage or in other countries can change a rank even when a country's own work-time value has not changed. A common price reduction lowers all countries' hours proportionally without changing ranks within an unchanged sample.
How can the hours fall?
Higher minimum wages are one route to greater AI purchasing power. The exchange rate supplies another: fewer local currency units per dollar make the same dollar price cheaper in local working time. Wage increases and currency depreciation can offset each other. The formula describes these relationships with the product and wage population held fixed. A statutory increase lowers the benchmark mechanically. For it to improve a worker's actual purchasing power, the wage needs to apply and be paid; coverage and compliance matter.
| Change | What IH records |
|---|---|
| Higher applicable local hourly wage | Fewer hours at that wage floor. |
| Fewer local currency units per USD | Fewer hours through currency appreciation against the dollar. |
| Lower common USD subscription price | Fewer hours everywhere; ranks stay the same in an unchanged sample. |
| Higher labor productivity | Only the effects transmitted to the applicable wage floor or exchange rate. |
Productivity is not an additional term in the headline formula. If it rises without changing the wage floor or exchange rate, IH stays the same. This makes the transmission of economic gains a question to investigate. The ILO's wage-setting guide considers productivity and sharing the gains of progress. A stronger currency is not a comprehensive measure of national competitiveness or an automatic consequence of higher productivity; these arithmetic channels are not a causal policy model.
Domestic AI supply and regional pricing can reduce actual access costs, but they require their own product, price and capability comparisons. The common US-price benchmark remains fixed in scope so that these alternatives do not silently change what the headline measures.
The benchmark basket.
The index basket contains ChatGPT Plus, Claude Pro and Google AI Pro, with equal weight in a median. All three approved prices must be present. An absent price never turns a three-plan basket into a two-plan basket.
Only regular monthly web subscriptions with a published official USD price qualify. Annual-payment equivalents, introductory discounts, trials, app-store prices, API usage and business plans are excluded. Taxes and local checkout prices are outside the headline measure. Other official consumer plans are available as individual comparisons.
The basket is a recognizable affordability benchmark, not a claim that the products offer identical capabilities. Changing its membership requires a new version and a published rationale. Past versions are preserved.
These named plans define the basket. They are not a guarantee of access to every provider's most capable model, every feature or unlimited usage. We use subscriptions because they are recognizable consumer purchases; a stable monthly price can still buy a changing service. Measuring the price of a fixed amount of model performance would require a separate quality-adjusted series.
The historical 2023–2025 paper uses a separate, explicit USD 20 benchmark. It does not impute a
three-provider basket before all three qualifying subscriptions existed. The historical
benchmark and the current complete basket-v1 answer related questions with different
product rules.
Whose minimum wage?
Where a published wage is available but an hourly normalization or exact regional match still requires an assumption, the calculator shows an estimated reference with the amount, working-time divisor, scope and source dates beside the result. These values remain outside the verified ranking. An assumed 40-hour week is labeled explicitly; it is not silently presented as local law. National-floor references also identify higher regional or sectoral rates.
Comparison-only monthly prices can be selected as price scenarios. They are excluded from basket-v1, the all-paid-plans median, the highest verified price and recorded official price snapshots. API and CSV results preserve the scenario and wage-reference labels; they do not represent a purchase offer. OpenAI Pro 500 is separate: it has been registered as an official $500/month plan since September 29, 2026, based on OpenAI's Pro-tier documentation and DevDay announcement.
The denominator is the general adult, unskilled, non-tipped statutory wage. Where occupation or employer coverage is essential, the population is stated alongside the observation. Special youth, apprentice and tipped rates are not a substitute for the general adult rate.
Countries open at the jurisdiction containing their capital. The label always includes that jurisdiction. A capital observation is not a national average or a population-weighted wage. Geography on the map is for visualization only.
For the United States, the scope is the higher general federal/state floor for FLSA-covered work, including geographic state wage zones. Municipal and county ordinances are not comprehensively covered. Canada uses provincial/territorial general rates; the federal industry-specific rate is separate. Japan uses the prefectural general rate.
Where the official wage is hourly, we use it directly. A daily, weekly or monthly rate needs an evidenced working-time conversion. We do not divide all monthly wages by a universal 174 hours. Paid rest, statutory bonuses and the annual number of wage payments must be documented before comparison.
A general national minimum can have limited coverage. For example, an Australian national rate applies to award- and agreement-free employees. Sectoral or collective rates must be labeled with their actual population.
The minimum wage is a legal benchmark, not the average price of human labor. It makes the comparison concrete for workers at the applicable wage floor, while leaving enforcement and actual earnings as separate questions. For a worker paid below that floor, the same subscription requires more hours. Establishing how common that situation is requires current earnings and compliance evidence, not the statutory rate alone.
Taxes and take-home pay change the question.
The headline divides a pre-tax US list price by a gross minimum-pay reference. A consumer-tax
sensitivity instead multiplies that price by 1 + tax rate; a take-home-pay
sensitivity divides by the employee's net hourly pay. These are separate adjustments. Standard
VAT or GST rates describe an explicit scenario, not a verified local checkout, and exemptions,
seller collection rules and subnational taxes can change the amount paid. The country
registers document local-language laws, effective dates and payroll conditions. Canada and the
United States use federal-only references; they do not capture all subnational consumer taxes.
The 2025 pre-tax Luxembourg–Brazil comparison is 1.14 versus 16.20 hours, a 14.18-fold spread. Luxembourg's 17% standard VAT scenario gives 1.34 hours. Brazil has documented municipal and federal service-tax provisions; their combined incidence depends on the supplier route, municipality and tax base. A single all-in subscription rate is not assumed. Separate ISS-only scenarios are provided with their excluded federal and payment-route charges.
Across the 35 countries with a documented standard-rate scenario, tax-adjusted hours range from 1.34 in Luxembourg to 15.50 in Peru. This available-rate subset excludes Brazil and cannot replace the 36-country pre-tax headline range.
Gross wages, net wages and employer costs serve three different decisions. Gross statutory wages provide a transparent legal benchmark. Net pay is relevant to a worker's disposable-income purchase; gross pay plus employer contributions is relevant to an employer's labor-cost comparison. Employee taxes depend on earnings, household circumstances and tax year, so an average-worker tax wedge cannot silently replace minimum-wage take-home pay. The paper does not report an observed net-wage Luxembourg–Brazil ratio without a specified household and tax basis.
Local-language official sources support explicit minimum-wage payroll examples. In Brazil, the 2025 minimum is R$1,518, the employee INSS deduction is 7.5%, and this ordinary monthly income is below the personal-income-tax threshold. The modeled net wage is R$1,404.15, giving a net/gross factor of 0.925. Applying that factor to the comparative gross-hourly benchmark gives 17.51 pre-tax subscription hours in a selected net-pay scenario. Other income, voluntary deductions, transfers and extra salary payments are outside that ordinary-month example. Luxembourg's selected class-1 regular-payroll model gives a net/gross factor of 0.88114; its application to the comparative hourly benchmark gives 1.30 hours before VAT, or 1.52 with its 17% VAT scenario. These applications are not exact matched payroll observations. The selected-profile pre-tax Brazil/Luxembourg ratio is 13.51. Applying only Brazil's ISS component of 2–5% gives 17.86–18.38 scenario hours and 11.78–12.13 times the Luxembourg VAT case; these partial-tax cases are not all-in checkout prices. The v5 supplement records the country sources and assumptions, and keeps the OECD's 2024 net-family-income cases separate from 2025 payroll.
Consumer taxes and employee deductions can lengthen hours more in some high-wage countries, while pay below the statutory floor can lengthen hours in low-compliance settings. Their net effect on the international spread is undetermined without matched evidence. Wage-floor comparisons also need attention to gender and coverage: they do not describe every worker's position in the earnings distribution. Read the tax and wage-definition sources.
Missing values mean different things.
- Estimated reference: a sourced wage with an explicitly stated conversion or scope limitation. Work time is shown, but it is not assigned a verified country rank.
- No general minimum: official evidence establishes that no general statutory floor applies. These observations are N/A and excluded from rankings.
- Awaiting verification: the applicable source, amount, geographic zone, population or hourly conversion is not yet established. This is unfinished research, not structural N/A.
- Collective reference: an explicitly identified agreement can provide context but is excluded from the statutory ranking.
- Unavailable price or FX: a required input is absent. Hours are
null, never zero.
Exchange rates and dates.
We use Frankfurter v2 reference rates based on central-bank and official institutional data. The quote is local currency per USD. The raw response and contributing providers are retained, including contributor dates. Blended reference rates may differ from bank, card or checkout rates.
A quote must have a reference date no later than the index date. On holidays we retain the most recent available reference. Prices are flagged stale after 7 days without verification; wages after 45 days; reference FX after 7 days. A collection failure leaves the last approved value intact with its original date.
Observations record collection, verification, approval and effective dates separately. A published future wage takes effect only from its effective date. Some official summary tables provide a reference date rather than the precise commencement of the underlying law; the evidence explicitly identifies that limitation.
History, corrections and editions.
Daily observations begin at the first verified publication. There is no synthetic back history. A correction creates a new observation and a new snapshot revision; the previous records and review trail remain available to editors. Dates are interpreted in UTC.
The public history returns the latest recorded revision for each actual observation date. The CSV and JSON expose methodology and basket versions, input values and source links. Recorded revision downloads retain the exact stored observation and its receipt ID, including earlier revisions when an ID has been retained. They do not recalculate past values from today's inputs. Formal monthly editions will be frozen manifests of snapshot IDs once coverage passes the release checklist, with annual editions for longer comparisons.
What the OECD data can tell us.
The OECD real minimum-wage series is valuable historical context. It uses constant-price wages. The current headline requires a current nominal wage to match a current nominal subscription price. Mixing a constant-price wage with today’s dollar price would create a misleading ratio.
Converting both the subscription price and the wage by the same purchasing-power factor cancels
that factor: (P / PPP) / (W / PPP) = P / W. It does not produce an independent “PPP
Intelligence Hours” measure. The paper's PPP robustness check instead substitutes a PPP rate for
the market exchange rate in the common USD benchmark. It asks about the domestic purchasing
power of the wage, a different question from the work needed to pay a dollar invoice.
Median-wage, after-tax and performance-adjusted comparisons are possible extensions to the current calculator, each needing its own population, reliable matched observations and versioned methodology. A published minimum-to-median wage ratio is not automatically a valid way to infer a current hourly median wage.
For a matched population and period, let κ be the Kaitz ratio: minimum wage divided by median
wage. Then IH_median = κ × IH_minimum. A median-wage comparison answers a different
question and may change country rankings. We do not infer a current hourly median from an older
national ratio or combine it with a regional minimum without a matching basis.
In the paper's 32-country scenario, rank correlations are ρ = 0.949 and τ = 0.823. The median absolute rank change is two places and the upper quartile three; the United States moves from 14th to 4th and the United Kingdom from 3rd to 8th. These are results of a κ-rescaling scenario, not independently observed hourly median wages. The published national earnings ratios may differ from the harmonized minimum-wage inputs in year, earnings concept or regional coverage; Chile has no same-year 2025 ratio and is excluded from this scenario. The replication retains those input definitions; no fully matched observed median-wage result is available.
Research background and interpretation.
Why measure a price in working time?
Labor-time prices have a long history. Cox and Alm's Time Well Spent (1997) expresses the cost of goods in the work needed to buy them. Ashenfelter's Comparing Real Wage Rates (2012) compares what workers doing standardized jobs can purchase across countries. Intelligence Hours applies this measurement tradition to paid AI access. It fixes a subscription benchmark; it does not hold workers' jobs, skills or productivity constant. Read the time-price references.
A Big Mac Index for the AI age.
The ambition is to use a recognizable purchase to make a global economic relationship worth watching, comparing and improving. The Economist's official data and methodology explain its burger-price comparison. IH uses a specified AI subscription and asks how much local labor can buy it. The analogy describes the public role of the index; its arithmetic answers a different question from purchasing-power parity or currency valuation.
The Big Mac literature shows why a recognizable product need not be a clean currency yardstick: local wages, rent, traded ingredients and other inputs all contribute to its price. Parsley and Wei (2007), Pakko and Pollard (1996, 2003), and Clements, Lan and Seah (2012) examine those differences and the limits of purchasing power parity comparisons.
IH makes a different choice: the headline holds the published US subscription price common across countries and asks how much local work buys it. With a common price, the ranking is exactly the inverse of the ranking of dollar-converted hourly wages. The contribution is a readable unit and a reproducible comparison. This choice neither establishes a worldwide checkout price nor assumes that AI has no local costs. IH is not an estimate of currency overvaluation or undervaluation.
Unequal wages are context, not a causal result.
Samuelson's factor-price equalization results describe conditions under which trade can equalize factor prices. Research by Trefler and by Clemens, Montenegro and Pritchett documents the importance of productivity, place and barriers when interpreting wage differences. This literature helps frame the question behind IH. A comparison of statutory wage floors is not itself a test of factor-price equalization, or evidence that AI will make wages converge. Read the trade and wage references.
A supplemental check of country-level Claude usage.
At a common price, IH is an inverse transform of the dollar minimum wage, which is strongly associated with national income. The exploratory Anthropic Economic Index check therefore asks a limited question: does IH add information beyond income about aggregate Claude usage? In the corrected 36-country 2025 sample, adding IH changes R² from 0.5867 to 0.5868 (ΔR² = 0.000095; partial r = −0.015; HC3 coefficient p = 0.926). Its outcome includes Free and Pro conversations relative to the working-age population; it is not a paid-subscription adoption rate or a measure of minimum-wage workers' use. The August 2025 sample and April–May 2026 robustness check use different observation windows, so the latter is not a contemporaneous replication. The small sample and shared income signal limit what an incremental association can establish. This is not a definitive discriminant-validity test or proof that affordability has no effect. Skills, organizational capacity and language are plausible complements to investigate, not effects estimated by IH. Read the supplemental inputs, design and estimates. View supplemental Figure D1.
The denominator changes the question.
Minimum-wage IH describes a legal wage floor; median-wage IH would describe a different position in the earnings distribution. The Kaitz identity above connects the two only when their inputs match; a rescaling with mismatched inputs is a scenario. Studies by Rani and colleagues and by Kanbur and Ronconi also show why legal coverage and enforcement matter. Neither a national average nor the midpoint of a regional wage range can silently replace the jurisdiction displayed in this index. Read the wage-measurement references.
Access costs and automation are different measurements.
The task-based work of Zeira and Acemoglu and Restrepo motivates interest in AI access, but IH does not measure an automation threshold or the machine-to-labor cost of a task. A subscription fee divided by a minimum wage does not measure the cost of completing the same task to the same standard. Productivity, usage limits, supervision, integration costs and the wage of the affected occupation all matter. Workplace productivity findings such as Brynjolfsson, Li and Raymond's cannot be applied as a universal conversion from subscription hours into labor saved.
Historical arguments about high wages and labor-saving innovation also depend on whose wages are measured, as the Allen–Humphries debate illustrates. IH can inform a task-specific cost comparison once those additional inputs are known; it does not predict job losses or adoption from the subscription price alone. Read the automation references.
Regional prices, product tiers and capacity.
A cheaper plan may differ in model access, usage or features. Deneckere and McAfee's work on quality differentiation offers one possible interpretation, not proof of a provider's motive. OpenAI's January 2026 Go announcement, for example, specifies an $8 US price and localized prices in some markets. Netflix's ₹149 Indian offer is a Mobile tier. Neither is evidence that an identical service has one price worldwide.
The paper's Plus pilot compares local listed prices with a pre-tax USD 20 US reference. On a consistent pre-tax basis, assuming the Indian and Indonesian lists include their applicable taxes, localization ratios become 0.97 and 0.95; Korea's reported VAT-inclusive list gives 0.93 after removing 10% VAT. India and Indonesia's tax inclusion is unverified. The pilot supports approximate pre-tax parity with modest discounts under these scenarios. This is a different scale from the Go listed-price ratios of about 0.57 in India, Indonesia and Nigeria, and the products differ. The US Go reference postdates the 2025 launches; continued local prices on its January 2026 introduction date have not been established.
Capacity constraints can affect prices and availability, as the peak-load pricing literature explains. A sales pause, however, is not an observed price increase, and an unannounced plan is not a price observation. A domestic compute strategy would also need its own cost and capability comparison: a long IH alone does not establish that producing AI locally would be cheaper. Read the pricing references and the national compute reference.
Reading changes over time.
For an unchanged basket, wage population and jurisdiction, the common-price measure has an exact logarithmic decomposition:
This separates subscription-price, exchange-rate and wage movements. Log changes add exactly; ordinary percentage changes do not. A carried-forward exchange rate gives a zero FX contribution by construction, not evidence that the currency was stable. Annual-average research panels also answer a different question from an observation on a particular date.
The paper's local-price extension uses IH_local = local price / local wage. If
expressed relative to the US benchmark, let L = local price / (USD price × FX). Its decomposition includes an additional Δ ln(L) term for changes in local pricing, taxes and other included charges. Merely adding
a standard tax rate to a US price is a scenario, not a verified local checkout observation.
Affordability is one part of digital access.
Research on telecommunications affordability and the digital divide distinguishes the cost of access from skills, use and outcomes. Language also matters: Ahia and colleagues find differences in token costs and utility across languages in API services. That does not directly measure subscription prices, but it helps explain why equal spending need not buy equal practical value.
The Broadband Commission's 2025 target puts entry-level broadband below 2% of monthly GNI per capita in low- and middle-income countries. That target has its own service and income denominator. Applying 2% to AI subscriptions or minimum-wage income would be a new policy choice, not an existing international standard. We publish hours without imposing an affordability pass/fail threshold. Read the digital-access references.
Research extensions.
Historical panels, median-wage comparisons, local checkout prices and performance-adjusted measures can extend this work. Each needs an explicit population, time basis, product definition and record of missing or carried-forward inputs. An annual national wage series is not interchangeable with the current capital-jurisdiction observation. Replacing a market exchange rate with a PPP factor prices a hypothetical purchasing-power basket; it does not convert the same dollar subscription into its actual local cost.
Research hypotheses about tier launch order, labor substitution or national compute policy need evidence beyond the index. Research tables and exploratory estimates do not become live observations until their underlying price, wage and exchange-rate inputs have passed the same source review as other published values.
Dollar-priced intelligence.
Author's interpretation and research hypothesis; not an empirical finding.
Dollar-priced AI gives workers another international price to watch. If AI purchases are invoiced and settled in dollars, wider access could add a channel of dollar demand; the outcome also depends on local-currency billing, settlement, hedging and the provider's own currency exposure. IH observes none of those flows and cannot establish reserve-currency effects or a government's pricing intent. The US AI export order provides policy context for the author's question. Section 5.4 of the public v5 paper keeps this interpretation separate from the wage, price and exchange-rate evidence.
What this index does not measure.
This is not a measure of model intelligence, capability, employment displacement or human worth. Nor is it a complete measure of access: devices, connectivity, payment methods, language, local availability, free tiers and usage limits all matter. Service availability shows consumer web access from dated official country and territory lists, with source links and review dates. It does not verify local paid-plan checkout or every included feature. “Country access unconfirmed” means we have not verified an official statement for that location; it does not mean the service is down. Regional exceptions, including Hong Kong and Macao, are recorded separately. A basket is available only when every constituent service is supported; any unsupported service makes it unsupported, and otherwise missing evidence keeps it unconfirmed. Price scenarios have no service availability.
The official status links on the index and country pages report outages and incidents separately from geographic access. We link to the providers' current reports without polling or copying a one-time operational status. Google's linked dashboard covers AI Studio, not the Gemini consumer app or all Google AI subscription features.
The “eight-hour day” is only a visual unit. The original total hours are always retained. Gross wages do not describe disposable income, actual earnings, compliance with wage laws or the hours a particular worker is offered.
Geographic coverage.
The catalog includes 38 OECD members plus India, China, Brazil, Indonesia, South Africa, Nigeria, Vietnam, Thailand, Malaysia, Philippines, Singapore and Saudi Arabia. Each wage observation identifies its jurisdiction and covered population. Country rankings use capital jurisdictions; the calculator also offers regional comparisons where data are available.
Research references.
These selected readings explain the measurement tradition, economic context and limitations. The notes describe their relevance to IH; they do not imply that the authors studied this index or endorsed its conclusions. References use journal issue years, with an earlier online year noted where relevant. Current prices and wages are supported by the official data sources below.
Labor time, living standards and the Big Mac comparison (7)
- The Economist. (n.d.). The Big Mac Index: Data and Methodology. Retrieved 1 October 2026. Original source
The primary data and calculation for the familiar-product comparison. The AI-age analogy concerns the public role of IH; its labor-time formula does not estimate currency misvaluation.
- Cox, W. M., & Alm, R. (1997). Time Well Spent: The Declining Real Cost of Living in America. Federal Reserve Bank of Dallas, Annual Report. Original source
A precedent for expressing the cost of goods in working time. Its US time series does not establish international comparability.
- Ashenfelter, O. (2012). Comparing Real Wage Rates. American Economic Review, 102(2), 617–642. DOI
Compares the purchasing power of workers doing standardized jobs. Intelligence Hours instead fixes a subscription benchmark and uses statutory wages with different coverage.
- Pakko, M. R., & Pollard, P. S. (1996). For Here or To Go? Purchasing Power Parity and the Big Mac. Federal Reserve Bank of St. Louis Review, 78(1). DOI
Explains why a familiar product can be a useful comparison without being a precise measure of currency misvaluation.
- Pakko, M. R., & Pollard, P. S. (2003). Burgernomics: A Big Mac Guide to Purchasing Power Parity. Federal Reserve Bank of St. Louis Review, 85(6), 9–28. DOI
Reviews non-traded inputs, trade barriers and other reasons for departures from purchasing power parity.
- Parsley, D. C., & Wei, S.-J. (2007). A Prism into the PPP Puzzles: The Micro-Foundations of Big Mac Real Exchange Rates. The Economic Journal, 117(523), 1336–1356. DOI
Separates the contributions of traded and non-traded inputs to Big Mac price differences. This motivates an explicit common-price benchmark; it does not establish identical AI costs worldwide.
- Clements, K. W., Lan, Y., & Seah, S. P. (2012; online 2010). The Big Mac Index Two Decades On: An Evaluation of Burgernomics. International Journal of Finance & Economics, 17(1), 31–60. DOI
Evaluates the information and biases in a single-product index. Simplicity still requires an explicit account of measurement limits.
International prices and unequal wages (5)
- Samuelson, P. A. (1948). International Trade and the Equalisation of Factor Prices. The Economic Journal, 58(230), 163–184. DOI
The theoretical benchmark for factor-price equalization, conditional on assumptions about technology and trade.
- Samuelson, P. A. (1949). International Factor-Price Equalisation Once Again. The Economic Journal, 59(234), 181–197. DOI
Refines the conditions for equalization. Differences between statutory minimum wages alone are not a test of those conditions.
- Trefler, D. (1993). International Factor Price Differences: Leontief Was Right! Journal of Political Economy, 101(6), 961–987. DOI
Shows the importance of international productivity and factor-price differences in interpreting trade data.
- Clemens, M. A., Montenegro, C. E., & Pritchett, L. (2019). The Place Premium: Bounding the Price Equivalent of Migration Barriers. Review of Economics and Statistics, 101(2), 201–213. DOI
Estimates wage gaps across locations while addressing worker comparability and selection. These estimates should not be inferred from an IH ranking.
- Cavallo, A., Neiman, B., & Rigobón, R. (2014). Currency Unions, Product Introductions, and the Real Exchange Rate. Quarterly Journal of Economics, 129(2), 529–595. DOI
Studies identical-product pricing across borders and currency areas. It does not establish that AI subscriptions have a universal checkout price.
Minimum wages, the Kaitz ratio and enforcement (8)
- International Labour Organization. (2015, December 3). 5.4 Economic Factors. Minimum Wage Policy Guide. Original source
Discusses productivity in minimum-wage setting and sharing economic gains. Productivity is not a direct input to IH; its effect must reach the applicable wage floor or exchange rate to alter the headline hours.
- OECD. Minimum Relative to Average Wages of Full-Time Workers. Earnings database, MIN2AVE. Original source
The source for minimum-to-median earnings ratios. The paper uses κ rescaling as a sensitivity scenario; differing years, earnings concepts or regional coverage prevent interpreting every rescaled value as an observed hourly median wage.
- Chile, Ministerio de Hacienda. (2025). Decreto Supremo No. 3, dated 8 January and published 6 February 2025. Original source
Records CLP 510,636 effective 1 January 2025. Together with the May change, it supports the public paper correction from frozen CLP 519,333 to a month-weighted CLP 522,878.6667. The dated conversion policy uses a 45-hour ordinary week through 25 April 2024 and 44 hours from 26 April 2024; these analytical hourly equivalents remain distinct from the notified monthly wage.
- Chile, Dirección del Trabajo. (2025, May 13). ORD. No. 322. Original source
Confirms on page 2 that Decreto Supremo No. 3 applies CLP 510,636 from 1 January 2025. Instrument, publication and effective dates are distinguished in the public correction provenance.
- Chile, Biblioteca del Congreso Nacional. (2025). Law No. 21.751, Article 1. Original source
Sets the adult statutory monthly minimum at CLP 529,000 effective 1 May 2025. The paper weights four months at CLP 510,636 and eight at CLP 529,000, preserving the original observations separately.
- Rani, U., & Belser, P. (2012). Low Pay among Wage Earners and the Self-Employed in India. International Labour Review, 151(3), 221–242. DOI
Documents low pay in India and motivates distinguishing legal wage floors from observed earnings.
- Rani, U., Belser, P., Oelz, M., & Ranjbar, S. (2013). Minimum Wage Coverage and Compliance in Developing Countries. International Labour Review, 152(3–4), 381–410. DOI
Examines coverage and compliance. A statutory benchmark does not describe the wages of every covered or informal worker.
- Kanbur, R., & Ronconi, L. (2018). Enforcement Matters: The Effective Regulation of Labour. International Labour Review, 157(3), 331–356. DOI
Distinguishes the law on the books from effective enforcement, a central limit of a minimum-wage denominator.
Dated wage units and worker coverage in the 2026 research snapshot (9)
- Japan, Ministry of Health, Labour and Welfare. (2025). Prefectural Minimum-Wage Deliberations and Effective Dates, FY2025. Original source
The effective-date table distinguishes rates already in force on 1 January 2026 from later increases. The fiscal-year weighted average does not establish a date-matched national point.
- Japan, Akita Labour Bureau. (2025, September 22). Revision of the Akita Prefectural Minimum Wage. Original source
States that JPY 1,031 takes effect on 31 March 2026 and that JPY 951 continues until 30 March. The January snapshot therefore retains JPY 951 as its prefectural lower endpoint.
- Poland, Council of Ministers. (2025, September 11). Minimum Remuneration and Minimum Hourly Rate in 2026. Dz.U. 2025, Item 1242. Original source
Sets PLN 4,806 monthly remuneration and the distinct PLN 31.40 hourly civil-contract floor from 1 January 2026. The paper normalizes ordinary monthly employment by 173⅓ analytical hours and preserves the civil-contract case separately.
- China, Xinjiang Human Resources and Social Security Department. (2024). Minimum-Wage Adjustment, Notice No. 66, effective 1 January 2025. Original source
The CNY 17.5 third-tier regional part-time rate includes employee contributions. It is a regional case, not a verified national minimum or an ordinary full-time hourly floor.
- China, Beijing Human Resources and Social Security Bureau. (2025). Minimum-Wage Adjustment, Notice No. 7, effective 1 September 2025. Original source
The CNY 27.7 regional part-time rate includes employer and employee pension, medical and unemployment contributions. Its contribution basis differs from Xinjiang's, so these cases do not establish a harmonized national range.
- India, Government of NCT of Delhi, Labour Department. (2025, April 15). Minimum-Wage Dearness-Allowance Order, effective 1 April 2025. Original source
Reports INR 710 daily and INR 18,456 monthly for the regional unskilled category. The paper retains a dated illustration with an explicit eight-hour assumption; exact continuation at the January 2026 cutoff is not independently established.
- India, Ministry of Labour and Employment. (2024, February 8). National Floor Level Minimum Wage. Rajya Sabha Unstarred Question No. 735. Original source
Describes the historical INR 176 daily floor as non-statutory and dating from 2017. It does not verify the copied INR 178 reference or a national statutory January 2026 range.
- Nigeria, Presidency. (2024, August 4). Presidential Broadcast on the Nationwide Protest, paragraph 26. Original source
Corroborates the signed NGN 70,000 monthly minimum. A 40-hour-week conversion is analytical; the amount alone does not establish universal coverage or a notified national hourly tariff.
- Nigeria, Presidency. (2025, February 28). Statement on the 2025 Budget of Restoration. Original source
Provides subsequent official corroboration of NGN 70,000. The replication keeps the unresolved consolidated statutory coverage and exemption limits explicit.
Consumer taxes, take-home pay and employer costs (3)
- OECD. (2024). Consumption Tax Trends 2024. Original source
Documents VAT/GST systems and standard rates. Applying a standard rate to the common US list price creates a tax scenario; it does not verify a provider checkout or a consumer-specific tax liability.
- OECD. (2025). Taxing Wages 2025. Original source
Distinguishes employee income taxes and social contributions, employer contributions, and household benefits across specified earnings and household cases. An average-worker tax wedge is not a matched estimate of minimum-wage take-home pay.
- OECD. Calculator of Taxes and Benefits: TaxBEN. Original source
Provides official minimum-income cases. The retrieved 2024 net-family-income series includes specified benefits and regional conditions; it is preserved separately from the v5 selected 2025 legal payroll sensitivities, including Brazil and Luxembourg.
Automation, tasks and productivity (6)
- Zeira, J. (1998). Workers, Machines, and Economic Growth. Quarterly Journal of Economics, 113(4), 1091–1117. DOI
Connects technology adoption to relative production costs. Subscription prices alone do not identify task-level costs.
- Acemoglu, D., & Restrepo, P. (2018). The Race between Man and Machine: Implications of Technology for Growth, Factor Shares, and Employment. American Economic Review, 108(6), 1488–1542. DOI
Models automation and new tasks jointly. IH measures neither displacement nor the creation of new work.
- Acemoglu, D., & Restrepo, P. (2019). Automation and New Tasks: How Technology Displaces and Reinstates Labor. Journal of Economic Perspectives, 33(2), 3–30. DOI
Explains why task productivity and the allocation of work matter alongside factor prices when assessing automation.
- Allen, R. C. (2015; online 2014). The High Wage Economy and the Industrial Revolution: A Restatement. Economic History Review, 68(1), 1–22. DOI
Presents the historical argument linking relative wages and energy costs to innovation; read alongside the critique below.
- Humphries, J. (2013; online 2012). The Lure of Aggregates and the Pitfalls of the Patriarchal Perspective: A Critique of the High Wage Economy Interpretation of the British Industrial Revolution. Economic History Review, 66(3), 693–714. DOI
Shows why the workers represented by a wage series matter for an argument about labor costs and living standards.
- Brynjolfsson, E., Li, D., & Raymond, L. R. (2025). Generative AI at Work. Quarterly Journal of Economics, 140(2), 889–942. DOI
Studies productivity in a particular workplace setting. Its effects are not a universal multiplier for converting IH into labor saved.
Product tiers, capacity and observed prices (6)
- Deneckere, R. J., & McAfee, R. P. (1996). Damaged Goods. Journal of Economics & Management Strategy, 5(2), 149–174. DOI
A theory of quality differentiation and price discrimination. It offers an interpretation of tiers, not evidence that a particular AI service was deliberately degraded.
- Steiner, P. O. (1957). Peak Loads and Efficient Pricing. Quarterly Journal of Economics, 71(4), 585–610. DOI
Explains pricing under capacity constraints. A subscription price or sales pause alone does not measure the scarcity cost of compute.
- Levy, D., & Young, A. T. (2004). “The Real Thing”: Nominal Price Rigidity of the Nickel Coke, 1886–1959. Journal of Money, Credit and Banking, 36(4), 765–799. DOI
A historical example of a stable nominal price. The mechanisms behind it are not automatically explanations of AI subscription pricing.
- Cavallo, A., & Rigobón, R. (2016). The Billion Prices Project: Using Online Prices for Measurement and Research. Journal of Economic Perspectives, 30(2), 151–178. DOI
A methodological reference for collecting comparable posted prices if a local-checkout series is developed.
- OpenAI (16 January 2026). Introducing ChatGPT Go, Now Available Worldwide. Original source
Documents a lower-priced tier, differences in access and usage, and localized pricing. The stated $8 monthly price is the US price.
- Netflix (14 December 2021). Netflix for All. Now at Happy New Prices. Original source
Documents the ₹149 Mobile plan in India. It is an example of both tiering and local pricing, not an identical-product price comparison.
Affordability, digital access and compute (8)
- Hilbert, M. (2010). When Is Cheap, Cheap Enough to Bridge the Digital Divide? Modeling Income Related Structural Challenges of Technology Diffusion in Latin America. World Development, 38(5), 756–770. DOI
Examines affordability and income distribution in technology diffusion. An identical nominal price need not imply comparable access.
- Humlum, A., & Vestergaard, E. (2025). The Unequal Adoption of ChatGPT Exacerbates Existing Inequalities among Workers. Proceedings of the National Academy of Sciences, 122(1), e2414972121. DOI
Documents differences in adoption among workers. Subscription affordability alone does not explain adoption or realized benefits.
- Stone, M. E. (2006). What Is Housing Affordability? The Case for the Residual Income Approach. Housing Policy Debate, 17(1), 151–184. DOI
Provides a conceptual comparison for affordability measurement: a simple expenditure ratio does not account for all essential household costs.
- Broadband Commission for Sustainable Development. 2025 Advocacy Target 2: Make Broadband Affordable. Original source
Defines a broadband target below 2% of monthly GNI per capita in low- and middle-income countries. This is not an established AI or minimum-wage threshold.
- Barrantes, R., & Galperin, H. (2008). Can the Poor Afford Mobile Telephony? Evidence from Latin America. Telecommunications Policy, 32(8), 521–530. DOI
A precedent for comparing a digital-service basket with income and poverty measures rather than relying on nominal prices alone.
- van Dijk, J. A. G. M. (2006). Digital Divide Research, Achievements and Shortcomings. Poetics, 34(4–5), 221–235. DOI
Distinguishes material access, skills and usage. A price-to-wage ratio covers only part of access to a digital service.
- Ahia, O., et al. (2023). Do All Languages Cost the Same? Tokenization in the Era of Commercial Language Models. EMNLP, 9904–9923. DOI
Examines differences in API token costs and utility across languages. It motivates capability-sensitive research but is not a measure of subscription checkout costs.
- OECD (2023). A Blueprint for Building National Compute Capacity for Artificial Intelligence. OECD Digital Economy Papers, No. 350. DOI
Provides context on national compute capacity. A long IH does not by itself show that domestic infrastructure would be cheaper or more effective.
Official data sources.
Provider pricing and support pages establish prices. Government labor sources establish wage inputs; OECD and Eurostat can cross-check coverage and definitions. Unclear or blocked source pages become review items rather than automatically approved numbers.
- The National Minimum Wage in 2026
UK Low Pay Commission · wage - 2026 Minimum Wage
Republic of Korea Minimum Wage Commission · wage - National minimum wage — award and agreement free
Australia Fair Work Ombudsman · wage - Minimum wage rates and types
Employment New Zealand · wage - The minimum wage — questions and answers
German Federal Ministry of Labour and Social Affairs · wage - Le SMIC — gross statutory rate
French Ministry of Labour · wage - National Minimum Wage Order 2025 — effective January 2026
Irish Statute Book · wage - Minimum wage amounts
Government of the Netherlands · wage - Minimum wage statistics — July 2026
Eurostat · wage - Minimum wage — sectors and coverage
Norwegian Labour Inspection Authority · wage - Salary and collective wage agreements
Government of Iceland · wage - Does the Employment Act regulate the amount of salary?
Singapore Ministry of Manpower · wage - ChatGPT pricing
OpenAI · price - Claude pricing
Anthropic · price - Google AI plans — United States
Google · price - Grok pricing
xAI · price - Kimi membership — monthly billing
Moonshot AI · price - Frankfurter v2 reference exchange rates
Frankfurter / contributing official institutions · fx - Consolidated State Minimum Wage Table
U.S. Department of Labor · wage - Prefectural minimum wages and effective dates
Japan Ministry of Health, Labour and Welfare · wage - Current and forthcoming general minimum wages
Employment and Social Development Canada · wage - Minimum Wage Database — source data
Employment and Social Development Canada · wage - Minimum wages at current prices
OECD · methodology - Real minimum wages — constant-price context
OECD · methodology - Minimum wage in 2026
Czech Ministry of Labour and Social Affairs · wage - Minimum wage rates and effective dates
Estonian Tax and Customs Board · wage - Key social indicators
Lithuanian State Social Insurance Fund Board · wage - Social minimum wage
Luxembourg Labour and Mines Inspectorate · wage - Minimum wage in 2026
Slovak Ministry of Labour, Social Affairs and Family · wage - National Minimum Wage Amendment 2026, Gazette 54075
South Africa Department of Employment and Labour · wage - Decree 12,797 of December 23, 2025 — national minimum wage
Presidency of Brazil · wage - Government Decree 426/2025 — minimum wage for 2026
Hungarian National Legislation Database · wage - Minimum wage — April 2026, 182-hour basis
Israel National Insurance Institute · wage - Minimum Wages Order 2024 — rates, commencement and coverage
Malaysia Department of Labour Sabah · wage - Wage Committee Notification No. 14 — rates and eight-hour ordinary day
Thailand Ministry of Labour · wage - Current minimum wage notification register
Thailand Ministry of Labour · wage - Wage thresholds for counting Saudi employees in Nitaqat
Saudi Ministry of Human Resources and Social Development · wage - Labour Law — wages, Articles 89–95
Saudi Ministry of Human Resources and Social Development · wage - Minimum wages and wage protection in the Arab States
International Labour Organization · wage - Statutory Minimum Wage — May 2026
Hong Kong Labour Department · wage - Minimum Wage for Employees — January 2026 rates
Macao Labour Affairs Bureau · wage - Guaranteed monthly minimum wage for 2026
Portugal Directorate-General for Employment and Labour Relations · wage - Azores regional minimum wage — January 2026
Government of the Azores · wage - Madeira Legislative Decree 1/2026/M
Diário da República · wage - Labour Code — Article 271 hourly remuneration formula
Diário da República · wage - Working in Portugal — ordinary working hours
Government of Portugal · wage - Decree 293/2025/ND-CP — hourly minima and all 34 provincial wage-zone schedules
Government of Vietnam · wage - National Capital Region — current wage rates and effective dates
Philippines National Wages and Productivity Commission · wage - Wage Order NCR-28 — rates, coverage and eight-hour daily basis
Regional Tripartite Wages and Productivity Board — NCR · wage - National minimum wage signed into law — NGN 70,000 per month
Nigeria State House · wage - Labour Act, section 13 — hours are agreed, not a universal hourly divisor
Nigeria National Electronic Labour Exchange · wage - Monthly minimum wages, national currency, 2026-S2
Eurostat · wage - Country-specific minimum-wage coverage and working-time definitions, July 2026
Eurostat · wage - January 2026 monthly minimum income
ChileAtiende · wage - Reduction to 42 weekly hours from April 26, 2026
Chile Labour Directorate · wage - 2026 minimum salary and working-time reduction
Fondo Colombia en Paz · wage - 2026 minimum wages by occupation
Costa Rica Ministry of Labour and Social Security · wage - Beijing minimum-wage notice, effective September 2025
Beijing Municipal Government · wage - Current minimum-wage orders
Delhi Labour Department · wage - Delhi unskilled daily and monthly wages, April 2025
Delhi Labour Department · wage - Jakarta provincial minimum wage 2026
Jakarta Provincial Government · wage - Minimum wages in the general and northern border areas, 2026
Mexico Official Gazette · wage - National, cantonal and collective minimum-wage coverage
Swiss Confederation · wage - Geneva gross hourly minimum wage, 2026
Swiss Federal Department of Foreign Affairs · wage - Basel-Stadt minimum wage and supplements, 2026
Canton of Basel-Stadt · wage - Neuchatel minimum wage, 2026
Canton of Neuchatel · wage - About ChatGPT Pro tiers
OpenAI Help Center · price - DevDay 2026 Recap
OpenAI · price - ChatGPT supported countries
OpenAI · availability - Where can I access Claude?
Anthropic · availability - Where you can use the Gemini web app
Google · availability
Use the data.
Use the index provides downloads, a citation builder, country and plan codes, and examples for using the public API. JSON index and CSV export use the same calculation engine. The official current capital rankings preserve the site's eligibility rules and exact ties, so consumers do not need to reconstruct the ranking. Historical dates require a recorded snapshot; the history API returns each recorded day's latest revision.
Suggested citation: Intelligence Hours Index, methodology 1.0, jurisdiction, observation date, revision or retrieval date, intelligence-hours.org. Always retain the basis and regional label when republishing a result.
Personal wage calculations use your gross hourly wage in USD and run in your browser. They do not change the published index and are not included in shared URLs, API responses or exports.
On a first visit, the host's approximate country code selects a country when it is in the catalog. Otherwise, the United States, South Korea or Japan is selected at random. An explicit link takes priority, followed by a valid previous selection saved in this browser. No precise location permission is requested.