What is the difference between alpha-2 and alpha-3 country codes?

Alpha-2 codes are two letters (GB, US) and are the most common, used in domains and locales. Alpha-3 codes are three letters (GBR, USA) and are more human-readable and collision-resistant, used in passports and some standards.

When should I use the numeric ISO 3166-1 code?

The three-digit numeric code (826 for the UK) is language-independent and script-independent, which makes it ideal for systems that must avoid Latin letters, such as some financial messaging and statistical datasets.

Is the alpha-2 code the same as a country's top-level domain?

Usually but not always. Most ccTLDs match the ISO 3166-1 alpha-2 code (.de for Germany), but there are exceptions like .uk (ISO code GB) and .eu, which is not a country code at all.

Do ISO 3166-1 codes ever change?

Yes. Codes are reassigned when countries form, dissolve or rename. Retired codes (like CS for Serbia and Montenegro) are reserved and not immediately reused, so historical data should be interpreted with the date in mind.

How are region groupings determined?

The region shown here follows the UN M.49 standard continental grouping (Africa, Americas, Asia, Europe, Oceania). M.49 also defines finer sub-regions, but the continental level is what most applications use for high-level segmentation.

Which codes are safe to use for test data or internal placeholders?

Use the user-assigned ranges ISO promises never to allocate: alpha-2 codes AA, QM–QZ, XA–XZ, and ZZ; alpha-3 equivalents like XAA–XZZ; and numeric 900–999. Kosovo's conventional XK sits in this space, and ZZ is a common 'unknown country' sentinel.

Why does Armenia's numeric code sometimes appear as 51 instead of 051?

Because a system stored the numeric code as an integer, which strips the leading zero. ISO 3166-1 numeric codes are three-character zero-padded strings (051, 004, 008); storing them as numbers breaks lookups against properly padded data.

What is the ISO 3166-1 Country Codes?

Searchable ISO 3166-1 country code reference: two-letter alpha-2, three-letter alpha-3 and three-digit numeric codes for each country, with UN region grouping and worked examples like GB, GBR and 826. It runs free in your browser on Gera Tools, with nothing uploaded.

ISO 3166-1 Country Codes

Name: ISO 3166-1 Country Codes
Creator: Gera Tools
License: https://creativecommons.org/licenses/by/4.0/

ISO 3166-1 country code reference

ISO 3166-1 assigns each country and dependent territory three standardized codes: a two-letter alpha-2, a three-letter alpha-3, and a three-digit numeric code. These codes power locale identifiers, ccTLDs, currency tables, shipping systems and statistical datasets. This reference lets you search by country name or any code form and shows all three side by side.

How the three ISO code forms relate

Each country has one of each code type, kept in sync by the ISO 3166 Maintenance Agency:

Alpha-2 — two uppercase letters (GB, US, AM). Most widely used; the basis for ccTLDs and BCP 47 region subtags.
Alpha-3 — three uppercase letters (GBR, USA, ARM). More mnemonic and less prone to collision; used in passports and ISO 4217 currency codes (the first two letters of a currency code are the country’s alpha-2).
Numeric — three digits (826, 840, 051). Identical to the UN M.49 country code; script-neutral, which is useful where letters cannot be used.

To convert between forms you simply look the country up in the registry — there is no algorithm, the mapping is a fixed table. The region column groups countries by UN M.49 continental area for coarse segmentation.

Where each code type gets used in practice

Knowing which code form a specific system expects saves integration headaches:

System or standard	Code type used
HTML `lang` attribute, BCP 47 locale	Alpha-2 region subtag (`en-GB`)
Country code top-level domain	Alpha-2 (mostly; see exceptions)
Passport Machine Readable Zone	Alpha-3
ISO 4217 currency codes	First two chars of currency code = alpha-2 of issuing country
UN statistical databases, M.49	Numeric
Financial messaging (SWIFT, ISO 20022)	Alpha-2 or Alpha-3 depending on field
IANA language subtag registry	Alpha-2

Known exceptions and edge cases

A handful of codes trip up developers because the ISO registration and the real-world usage diverge:

United Kingdom is GB / GBR / 826 in ISO but .uk as a country domain (the .gb ccTLD was registered but has never been widely used).
Switzerland is CH / CHE / 756, from the Latin Confoederatio Helvetica — nothing to do with the English name.
Greece is GR but .gr — straightforward, but the alpha-3 is GRC, not GRE.
Kosovo has no ISO 3166-1 entry as of recent years (recognition is not universal). It is commonly assigned XK / XKX as a user-assigned code, not an official ISO one.
Timor-Leste is TL in ISO, not TP (the old Portuguese Timor code).
Retired codes like CS (Serbia and Montenegro) or YU (Yugoslavia) are reserved and will not be reused, but can still appear in old data.

Reading one country across all three forms

The United Kingdom illustrates every trap at once: ISO assigns GB / GBR / 826, yet the domain is .uk, licence plates use UK (since 2021), and UK itself is an exceptionally reserved ISO code — reserved so nobody else can take it, but not the official assignment. Armenia shows the numeric-code gotcha: its numeric code is 051, and systems that store numeric codes as integers silently turn it into 51, which then fails string comparisons. Store numeric ISO codes as zero-padded three-character strings, never as numbers.

Codes you can use without colliding with ISO

ISO 3166-1 deliberately sets aside user-assigned code space that the Maintenance Agency promises never to allocate. If you need internal placeholders (test data, “unknown country”, multi-country groupings), draw from these ranges and nowhere else:

Form	User-assigned range
Alpha-2	`AA`, `QM`–`QZ`, `XA`–`XZ`, `ZZ`
Alpha-3	`AAA`–`AAZ`, `QMA`–`QZZ`, `XAA`–`XZZ`, `ZZA`–`ZZZ`
Numeric	`900`–`999`

This is exactly why Kosovo is conventionally XK: it sits in the user-assigned XA–XZ block, so the (unofficial) convention cannot collide with a future official assignment. The same logic makes ZZ a common “unknown/unspecified” sentinel in address and payment systems.

Separately, ISO keeps a small set of exceptionally reserved codes that appear in the wild but are not country assignments: EU (used in ISO 4217 for the euro and in TIN/VAT contexts), UK (reserved for the United Kingdom), SU (legacy Soviet Union, still seen in the .su ccTLD), and FX (historical “metropolitan France”). Accept them where your upstream data uses them, but map them explicitly rather than treating them as ordinary entries.

Tips and notes

The alpha-2 code is your safest default for web and locale work because it matches BCP 47 region subtags and most ccTLDs.
For language tags, combine a language subtag with the alpha-2 region: en-GB, hy-AM, pt-BR.
When building a shipping or payment form, confirm whether the API expects alpha-2 or numeric — both are common and confusing them causes silent validation failures.
Reserved or transitional codes (EU, UK, XK) appear in many datasets; treat anything outside the official list as user-defined or exceptional, and document the choice explicitly.

Sources and references

ISO — Country Codes (ISO 3166) — the authoritative alpha-2, alpha-3, and numeric assignments
ISO 3166 Maintenance Agency — Online Browsing Platform — searchable official code list, including reserved/retired codes
UN Statistics Division — Standard country codes (M.49) — the numeric codes and region groupings
IANA Language Subtag Registry — how alpha-2 codes are used as BCP 47 region subtags

Maintained by the Gera Tools editorial team. Codes are reassigned when countries form, dissolve, or rename; treat historical data with its date in mind and confirm current assignments on the ISO OBP. Last reviewed 2026-07-02.