Military Rankings Methodology
Methodology version 3 — 29 August 2026.
GlobalMilitary.net ranks the world's militaries with four indexes: the composite ATLAS Index (Aggregate Total of Land, Air and Sea strength) and three single-domain indexes — AIRS (Air Inventory Rating of Strength) for air forces, SEAS (Sea Equipment Aggregate Strength) for navies and ARMS (Armour and Manpower Rating of Strength) for ground forces. All four are computed from our own equipment inventories — every aircraft type, ship class and vehicle line a country fields, weighted by what it is and how old it is — plus published figures for personnel, budgets and nuclear arsenals. Every weight and formula is on this page, so the score on any country page can be rebuilt from it and the inventory the page shows.
How to read a score. Each index runs from 0 to 100, and the strongest force sets the 100. The scales are relative, not percentages of capability: 50 does not mean half the leader's strength. The branch indexes take the square root of a force's share of the leader's points (50 = a quarter of the leader's inventory); ATLAS takes the fourth root of each pillar's share (50 = a sixteenth). So one point near the top is a large real-world gap, and one point near 20 a small one. Read the ranks, not the decimals: a data update on one country routinely moves its neighbours by a place.
Coverage. ATLAS ranks every sovereign state with armed forces for which we hold an inventory, a personnel figure or a budget — 170 countries in 2026. States without a military (Costa Rica, Panama, Iceland…) and territories keep their pages but are not ranked. AIRS ranks the 162 air forces with an active inventory, SEAS 80 navies and ARMS the 86 armies whose equipment we inventory line by line; a country missing from a branch ranking has no inventory there yet and reads zero on that ATLAS pillar (ground forces excepted — see below).
Older citations: the composite was the Global Military Index until 15 August 2026.
Our Principles
- Transparency. Weights and formulas are disclosed here and kept in step with the code that computes the live scores.
- Verifiable data. Scores come from our own equipment inventories and from published figures (CIA World Factbook, World Bank, SIPRI), never from opaque aggregates.
- Quality over raw counts. A 1960s fighter is not a modern one and a landing craft is not a destroyer: equipment is weighted by category and by generation, size or era.
- Honest limits. No index captures training, morale, logistics or alliances; the Limitations section says what ours do not measure.
The Four Indexes
| Index | Scope | Where to see it |
|---|---|---|
| ATLAS Index (Aggregate Total of Land, Air and Sea strength) | Overall military capability, six pillars | Strongest militaries ranking, country pages |
| AIRS Index (Air Inventory Rating of Strength) | Air power | Air forces ranking, air force pages |
| SEAS Index (Sea Equipment Aggregate Strength) | Naval power | Navies ranking, navy pages |
| ARMS Index (Armour and Manpower Rating of Strength) | Ground forces | Armies ranking, army pages |
The three branch indexes are standalone rankings. ATLAS reads its ground, naval and air pillars from them, on its own fourth-root curve — which is why a country's ATLAS air pillar and its AIRS score are different numbers built on the same points (France: air pillar 53.3, AIRS 27.5).
A few assets are deliberately read twice, each time for a different thing: the budget in the economic pillar and again in every availability factor; attack helicopters as airframes in AIRS and as a multiplier in ARMS; embarked fighters as airframes in AIRS and as the air-wing multiplier in SEAS; ballistic-missile submarines as hulls in SEAS and as a delivery mode in the nuclear pillar.
Every weight below is also a slider in the Build Your Own Ranking panel of the countries, air forces, navies and armies pages: move one and the same inventory is re-scored in your browser, next to the published ranking.
ATLAS Index (Aggregate Total of Land, Air and Sea strength)
Six weighted pillars, each scored as the fourth root of its share of the highest value observed worldwide — 100 × (value / max)^0.25 — so a country holding a sixteenth of the leader's value scores 50. The four conventional pillars are read after the availability factor of their branch. Arsenals span several orders of magnitude, and the composite adds six domains: a flatter curve pays a premium for breadth — on a log scale, thirty patrol boats were worth 40 of the 100 naval points — while a steeper one lets a single large inventory outweigh everything else.
| Step | Range | Full credit / basis | Applies to |
|---|---|---|---|
| Availability | ×0.80 → ×1.20 | the branch index's PPP budget per point of the force, on a wider band | each conventional pillar, before scoring |
| Pillar score | 0 → 100 | (value / max)^0.25 — the fourth root of the share of the world's highest value | each pillar |
| Calibration | 100 | the strongest military sets the scale; constants frozen between data updates | the total |
Manpower (10%)
active + 0.5 × reserve + 0.3 × paramilitary.
Ground, naval and air power (20% + 20% + 25%)
The inventory power of the ARMS, SEAS and AIRS indexes: the quality-weighted inventory with its capability multipliers (enablers and armament for aircraft; air wing, sustainment and armament for hulls; attack helicopters and armament for ground forces). Two terms of the branch scores are left out: the ARMS manpower term, a pillar of its own here, and the order books — ATLAS is a snapshot of forces in being. A country not yet covered by our army inventory gets a ground pillar imputed from its tank park (tanks × 10 × 1.52, the median ratio of inventory power to tank points among covered countries).
Nuclear deterrent (10%)
Warheads, log-scaled, then modulated by posture: ×0.8 with one delivery mode, ×0.9 with two, ×1.0 with a full air-sea-land triad. Host countries count the foreign warheads stationed on their soil at 15%; acquiring states get a small symbolic score.
Economic strength (15%)
Military expenditure (nominal USD) × a military purchasing-power factor — what a dollar of defense spending buys in soldiers, shells and steel in that economy, relative to the United States. For 59 countries the factor is Peter E. Robertson's published military PPP rate (The Real Military Balance, Review of Income and Wealth 68(3), 2022, doi 10.1111/roiw.12536; 2025 estimates from militaryppp.com): China 1.79, Russia 2.22, India 3.41 — and below 1.0 for Germany, Denmark or Australia, where world-price equipment and high pay mean a dollar buys less. The 21 countries the dataset does not cover (Pakistan, Iran, Israel, Saudi Arabia…) carry an editorial estimate on the same scale. The same factors feed the availability factors of the three branch indexes.
Availability
Serviceability rates are published for a handful of countries only, so each index uses a proxy: the PPP-adjusted budget per point of the force, log-mapped between the 10th and 90th percentiles of the field. The branch indexes map it onto ×0.90 → ×1.10; ATLAS, where it is the only reading of how well a force is funded, onto ×0.80 → ×1.20. The manpower pillar takes the ground force's factor. It is read per point rather than per soldier or per airframe so that nothing a country adds to its forces can ever lower its score — a per-soldier version rewarded demobilisation.
Final score
ATLAS = Σ weight × pillar score, rescaled so the strongest military scores exactly 100. Every country page shows the pillar-by-pillar breakdown.
A worked example: France, 2026
| Pillar | Input | After availability | Share of world's highest → score | Weight × score |
|---|---|---|---|---|
| Manpower | 270,000 active + ½ × 43,000 reserve + 0.3 × 100,000 gendarmerie = 321,500 | × 1.18 = 379,400 | 12.1 % → 59.0 | 10 % → 5.9 |
| Ground forces | 13,349 ARMS points (Leclerc, Jaguar, Griffon, Caesar…) | × 1.18 = 15,750 | 8.5 % → 54.0 | 20 % → 10.8 |
| Naval power | 2,593 SEAS points (carrier with 41 Rafale M, 4 SSBN, 5 SSN, 22 destroyers and frigates…) | × 0.98 = 2,540 | 15.8 % → 63.0 | 20 % → 12.6 |
| Air power | 16,029 AIRS points (264 combat aircraft, 14 tankers, 7 AEW&C, three armament cases) | × 1.04 = 16,670 | 8.1 % → 53.3 | 25 % → 13.3 |
| Nuclear deterrent | 370 warheads, air and sea (dyad × 0.9) | 370 | 6.8 % → 51.1 × 0.9 = 46.0 | 10 % → 4.6 |
| Defense budget | $66.4B × military PPP 1.38 | $91.6B | 10.0 % → 56.2 | 15 % → 8.4 |
The weighted sum is 55.6; the strongest military's is 98.8, which the calibration sets to 100 — France scores 56.3, sixth in the world.
How much the choices matter
For one change at a time, every other setting held at its published value: how many of the top 30 change rank, and the largest moves.
| Change | Top-30 ranks that move | Largest moves |
|---|---|---|
| Manpower 10 → 5 %, air 25 → 30 % | 16 | Vietnam 20 → 23, Bangladesh 34 → 41 |
| Manpower 10 → 15 %, budget 15 → 10 % | 17 | North Korea 17 → 14, Germany 13 → 16 |
| Nuclear pillar removed | 21 | North Korea 17 → 23, Israel 11 → 16, Saudi Arabia 12 → 9 |
| Budget pillar removed | 21 | Ukraine 15 → 20, North Korea 17 → 13 |
| Pillar curve 0.25 → 0.20 (flatter) | 10 | Ukraine 15 → 16, Japan 7 → 8 |
| Pillar curve 0.25 → 0.35 (steeper) | 10 | Ukraine 15 → 12, Israel 11 → 14 |
| Pillar curve 0.25 → 0.50 (square root, as the branch indexes) | 17 | Ukraine 15 → 10, Bangladesh 34 → 28 |
| Availability band ±0.20 → none | 16 | North Korea 17 → 14, Greece 23 → 20, Taiwan 14 → 12 |
| Availability band ±0.20 → ±0.30 | 14 | Egypt 18 → 20, Algeria 19 → 17 |
The top eight do not move under any of these changes: the order United States, China, Russia, India is data, not weighting. Below that, the nuclear and budget pillars matter most, and the curve exponent little. What no weighting can do is decide a close pair — Israel and Saudi Arabia, Taiwan and Ukraine, Egypt and Algeria sit within a point of each other and trade places with every data update.
AIRS Index (Air Inventory Rating of Strength)
Each country's latest active inventory, on a square-root scale: Index = √(Power Score / Max Power Score) × 100.
Power Score = (Σ aircraft × category weight × quality + combat mass × (enablers × armament − 1) + order book) × availability
| Step | Range | Full credit / basis | Applies to |
|---|---|---|---|
| Enablers | ×1.00 → ×1.30 | 1 tanker per 25 combat aircraft, 1 AEW&C per 40; EW and ISR aircraft | the combat mass |
| Armament | ×1.00 → ×1.20 | a modern BVR missile · an air-launched cruise missile · an anti-ship missile | the combat mass |
| Availability | ×0.90 → ×1.10 | PPP budget per point of the fleet, $1.1M → $17.5M | the whole score |
| Order book | + ¼ per airframe | firm orders, never more than ¼ of the fleet's points | added before availability |
Generation tier. Every combat aircraft and bomber carries a capability tier from an editorial table, anchored on the constant-dollar flyaway cost of the variant.
| Tier | Value | Examples |
|---|---|---|
| 5th generation | 1.00 | F-22, F-35, J-20 (Su-57, J-35: 0.90) |
| 4.5 generation | 0.70 | Rafale, Typhoon, F-16V, Su-35, F-15EX, Super Hornet, J-16 |
| 4th generation, modernised | 0.45–0.60 | F-16C/D, F-15C/E, Su-30SM, Mirage 2000-5, J-10C, Gripen C/D, Su-34 |
| 4th generation, original standard | 0.25–0.35 | F-16A/B, MiG-29, Su-27S, Mirage 2000C, Tornado, Su-24, Su-25 |
| 3rd generation | 0.12–0.15 | F-4, MiG-23, Mirage F1, Jaguar, J-8 |
| 2nd generation, armed trainers | 0.02–0.08 | MiG-21, F-5E, J-7, A-4, Alpha Jet, K-8 |
The tier is refined per inventory line by the variant it records (an F-16AM is not an F-16A) and, where the fleet standard is documented, per country; a type not yet tiered is scored conservatively from its service entry year. Helicopters carry an era factor (0.05 for pre-1960 designs to 1.0 from 2000), combat drones their unit price relative to the MQ-9 Reaper; transports, special-mission aircraft and trainers count their category weight.
Enablers (×1.00 → ×1.30). Tankers and AEW&C aircraft each earn up to 12% on the combat mass (full credit at one tanker per 25 combat aircraft, one AEW per 40), electronic-warfare and ISR aircraft up to 4% and 2%. Maritime patrol aircraft are naval capability and are not scored here.
Armament (×1.00 → ×1.20). Three cases from our missile inventories, a third each: a modern beyond-visual-range air-to-air missile (AMRAAM, Meteor, PL-15…), an air-launched cruise missile (Storm Shadow, JASSM, Kh-101…), a modern air-launched anti-ship missile (Exocet, Harpoon, YJ-83…).
Availability (×0.90 → ×1.10). PPP-adjusted budget per point of the fleet, about $1.1M to $17.5M per point in 2026 (see Availability).
Order book. Firm orders count a quarter of a delivered airframe, capped at a quarter of the active fleet's points.
Worked example: France, 2026. 264 combat aircraft at a mean tier of 0.54 make 7,160 combat points; helicopters, transports, special-mission aircraft, drones and trainers bring the fleet to 12,535. Fourteen tankers and seven AEW&C aircraft earn ×1.24 on the combat mass, the Meteor, Scalp and Exocet ×1.20 — an inventory power of 16,029, the figure the ATLAS air pillar reads. The order book adds 1,238, availability ×1.02: a Power Score of 17,598 against the United States' 232,705, √(17,598 / 232,705) × 100 = 27.5, eighth in the world.
SEAS Index (Sea Equipment Aggregate Strength)
Each country's latest active inventory, on the same square-root scale.
Power Score = (Σ hulls × category weight × size × era × propulsion [minor and amphibious vessels capped] + carrier points × (air wing − 1) + surface-combatant points × (sustainment − 1) + combatant points × (armament − 1) + pipeline) × availability
| Step | Range | Full credit / basis | Applies to |
|---|---|---|---|
| Size | √(t / median) | displacement over the median of the category worldwide | every hull |
| Nuclear propulsion | ×1.5 | nuclear-powered submarines | submarines |
| Minor-vessel cap | ≤ 30 % | patrol, coastal, mine-warfare, unmanned and amphibious hulls together | share of the combatant points (carriers to corvettes, submarines) |
| Air wing | ×0.40 → ×1.00 | 24 embarked fixed-wing aircraft per carrier | the carrier points |
| Sustainment | ×1.00 → ×1.20 | 1 replenishment ship per 8 surface combatants | the surface combatants |
| Armament | ×1.00 → ×1.20 | an anti-ship missile · a land-attack cruise missile · an area air-defence missile | the combatants |
| Availability | ×0.90 → ×1.10 | PPP budget per point of the fleet, $8.5M → $215M | the whole score |
| Pipeline | + ½ · + ¼ | hulls fitting out ½, on order ¼, never more than ¼ of the fleet's points | added before availability |
Hull value. Category weight × size × era × propulsion. Size is the square root of the class's displacement over the median of its category (a 2,000 t diesel-electric boat counts about half a 7,000 t nuclear one); era reads the class's commissioning year, 0.25 for 1965 or earlier to 1.0 from 2005; nuclear-powered submarines carry ×1.5. A hull whose class has no published displacement or year counts exactly its category weight.
Minor-vessel cap. Patrol, coastal, mine-warfare, unmanned and amphibious hulls count, but together never more than 30% of the combatant points — a navy is not its patrol boats, nor its landing ships.
Air wing (×0.40 → ×1.00 on the carrier points). A carrier scores by the fixed-wing aircraft its navy can embark (F-35B/C, Super Hornet, Rafale M, J-15…), read from our air force inventories: 40% of its points with none, full credit at 24 per carrier.
Sustainment (×1.00 → ×1.20 on the surface combatants). Underway-replenishment ships, civilian-manned hulls included, are what makes a fleet blue-water; full credit at one per 8 surface combatants.
Armament (×1.00 → ×1.20 on the combatants). Three cases, a third each: a modern ship-launched anti-ship missile (NSM, Harpoon, YJ-18…), a sea-launched land-attack cruise missile (Tomahawk, Kalibr, MdCN…), an area air-defence missile (Aster, SM-2/SM-6, HHQ-9…).
Availability (×0.90 → ×1.10). PPP-adjusted budget per point of the fleet, about $8.5M to $215M per point in 2026.
Pipeline. Hulls fitting out count half a delivered hull and firm orders a quarter, together capped at a quarter of the active fleet's points.
Navy pages also group fleets into capital ships (carriers, cruisers), major combatants (destroyers, submarines, frigates), light combatants (corvettes, amphibious ships) and minor vessels (patrol, coastal, mine warfare, unmanned).
ARMS Index (Armour and Manpower Rating of Strength)
The ground forces of every country whose army equipment we inventory (86 in 2026), on the same square-root scale; a country without an inventory is absent rather than estimated.
Power Score = (Σ (active + 0.3 × stored) × category weight × era × rotary × armament + manpower + order book) × availability
| Step | Range | Full credit / basis | Applies to |
|---|---|---|---|
| Storage | × 0.3 | a vehicle in storage counts 30 % of a fielded one | every category |
| Manpower | 350 × √(thousand troops) | active personnel, square-root return | added to the equipment points |
| Rotary support | ×1.00 → ×1.15 | 2 attack helicopters per 100 tanks and IFVs | the equipment points |
| Armament | ×1.00 → ×1.20 | a modern ATGM · an area air-defence system · a 300 km+ strike system | the equipment points |
| Availability | ×0.90 → ×1.10 | PPP budget per point of the force, $0.17M → $5.6M | the whole score |
| Order book | + ¼ per vehicle | never more than ¼ of the fielded points | added before availability |
Equipment. Category weight × an era factor from the type's service entry — 0.15 for 1965 or earlier (a T-55), 0.45 at 1985 (a Leopard 1), 1.0 from 2010 (a Leopard 2A7, a K2, a T-90M) — refined per inventory line by the variant it records (a T-72B3 line is a 2012 rebuild). Vehicles in storage count 30%; lines not yet linked to a vehicle type carry 0.6.
Manpower. 350 points per √(thousand active troops): a square-root return, so the marginal soldier of a 1.2-million army counts less than that of a 60,000 one. Manpower makes 10–30% of a typical top-25 score — 4% for the United States, about 40% for North Korea.
Rotary support (×1.00 → ×1.15). Attack helicopters (Apache, Mi-28, Ka-52, Tiger, Z-10…) multiply the equipment mass, full credit at 2 per 100 tanks and infantry fighting vehicles.
Armament (×1.00 → ×1.20). Three cases, a third each: a modern anti-tank guided missile (Javelin, Spike, Kornet…), a medium or long-range surface-to-air system (Patriot, S-400, SAMP/T…), a conventional land-based strike system of 300 km or more (ATACMS, Iskander, Hyunmoo…).
Availability (×0.90 → ×1.10). PPP-adjusted budget per point of the force — equipment with its multipliers plus manpower — about $0.17M to $5.6M per point in 2026.
Order book. Vehicles on order count a quarter of a delivered one, capped at a quarter of the fielded points.
Known limits. About a third of our inventory lines are not yet linked to a vehicle type page and carry the neutral era value. A vehicle issued to a formation, regular or reserve, is filed active; stored means depot stock with no assigned unit — where a source gives one storage figure for both, the index understates the army. And it measures mass and era, not what a force does with them: Israel, whose edge lies in training, C4ISR and air-land integration, ranks below Egypt on any reading of the two inventories, and we publish that rather than tune it away.
Validation
No external ranking publishes its inventories or its arithmetic, so there is nothing to fit against. The indexes are validated the other way round: against facts the ranking must not contradict, and against properties any index of strength must have.
Ordering constraints. After every data update, an audit checks pairs whose order no informed reader would dispute — United States above China above Russia above India; Ukraine above Qatar; Pakistan above Singapore; Egypt above the UAE; Israel above Egypt in the air; Italy above Indonesia at sea; South Korea above North Korea on the ground, and some thirty more. When a constraint breaks we look for the data error first and discuss the weights second; we never edit a country to pass. Two ARMS constraints fail today and are published as such: Israel above Egypt (a limit of a mass-and-era index) and Germany above Myanmar (types with no equipment page yet on both sides).
Monotonicity. No input a country can raise may lower its score. The audit raises troops, tanks and budget by 10% for the sixty strongest countries and checks that every score rises or holds.
Reproducibility. The scores are recomputed from the database on every deployment, the calibration constants are frozen and printed in the code, and a snapshot of the four rankings is kept at each methodology version, so a later edition can say which moves come from the data and which from the method.
Data Sources and Updates
- Equipment inventories — our own database, per aircraft type, ship class and vehicle category, updated continuously from official announcements, defense media and open-source reporting; fleet changes are logged in the air force and navy event trackers.
- Personnel — World Bank (armed forces personnel), CIA World Factbook and national sources.
- Defence budgets — CIA World Factbook (military expenditure as a share of GDP, GDP at official exchange rates), nominal US dollars; SIPRI as a cross-check. Converted to military purchasing power with Robertson's rates.
- Nuclear arsenals — Federation of American Scientists estimates, via Our World in Data for the historical series.
Scores are recalculated whenever the inventories change materially, and the normalization constants recalibrated after major updates so the strongest force in each domain stays at 100.
Limitations
A reproducible synthesis of measurable military capability — not a prediction of who would win a war. The indexes do not capture training, morale, doctrine and combat experience; logistics depth, ammunition stocks and industrial mobilization; alliances, geography and the will to fight; cyber, space and electronic warfare; or the true serviceability of listed equipment, which official figures overstate. Where data is uncertain — nuclear arsenals above all — we use the most widely accepted published estimates.
Citing Our Indexes
Cite freely with attribution to GlobalMilitary.net, a link to the page, the index named in full on first mention and the date consulted, since scores move with the inventories. Suggested form:
GlobalMilitary.net, ATLAS Index (Aggregate Total of Land, Air and Sea strength) — Strongest Militaries in the World, https://www.globalmilitary.net/rankings/countries/military-power/, accessed <date>
Corrections: [email protected]. Questions: [email protected].