There is no IELTS for Arabic. Here is what there is instead

English testing converged. Arabic fragmented

IELTS, TOEFL and the Cambridge suite compete on the same construct, report on comparable scales, and are read interchangeably by admissions offices on every continent. A candidate can prepare once and be understood everywhere. Nothing like that exists for Arabic.

Instead there are roughly a dozen serious instruments, spread across ten jurisdictions, built by a French cultural institute, a Saudi academy, a Turkish measurement agency, a British chartered body, Spain’s autonomous regions, an Uzbek assessment agency, the United States defence establishment and the United Nations Secretariat. They do not report on the same scale, they do not test the same skills, and in several cases they are not even measuring the same language.

This is usually described as a failure of coordination. It is better understood as the predictable result of three properties of Arabic and one property of the demand for it.

Diglossia decides what a test is allowed to measure

Arabic operates as a diglossic system. The high variety — Modern Standard Arabic, fuṣḥā — dominates formal writing, legal codification, print journalism, broadcast news and diplomacy. The regional vernaculars are the medium of spontaneous speech and domestic life, and they are not mutually intelligible with the high variety. Levantine Arabic alone has roughly sixty million speakers.

That leaves every test designer with a choice that has no clean answer. Test Modern Standard only, and an “oral proficiency” score measures performance in a register native speakers rarely use conversationally: the construct is clean and travels across borders, but its validity for real spoken interaction is weak. Test the vernaculars, and scores gain communicative validity while losing all comparability — a Levantine rating and an Egyptian rating are not the same certificate.

Almost the entire civilian market resolved this the same way. CIMA, Hamza, the Spanish EOI system, YDS, the ALPT, the UN examination and the ACTFL Arabic instruments all target Modern Standard. The one systematic exception is the United States defence and intelligence community, whose language centre maintains standardized batteries for Egyptian, Iraqi, Levantine, Algerian, Saudi, Sudanese and Yemeni Arabic alongside the standard variety.

The page does not give you the vowels, and the traditions split on that

Printed Arabic is normally unvocalized. Short vowels and case endings are absent from the page and have to be recovered through morphological decoding and syntactic inference, which makes reading Arabic a materially different task from reading a fully specified orthography. Testing traditions divide on whether to assess that decoding directly.

The communicative models present unvocalized authentic texts and assess grammar indirectly, through comprehension. CIMA’s candidate guide states plainly that its reading texts are unvocalized, in line with press practice, and that candidates are not required to vocalize their own writing. Pearson’s A Level specification takes the same line: vocalization appears in questions only where it is needed for clarity, students are not expected to use it, and misuse is not penalised.

The structural and translation-oriented models treat explicit vocalization and case-ending analysis as a direct measure of grammatical mastery. Türkiye’s YDS demands parsing of unvocalized syntax without supplying any diacritics, and the Russian orientalist tradition places vocalization and naḥw analysis at the centre of assessment.

Neither is wrong. But they produce certificates that mean different things, which is one reason a score from one tradition tells an admissions officer in the other very little.

Nobody buys an Arabic certificate the way students buy IELTS

This is the property that explains the shape of the market. Demand for Arabic certification is generated almost entirely by specific institutional mandates rather than by individual consumers: a civil-service pay tier, a university entrance waiver, a defence bonus, a translator’s professional registration. Each mandate produced an instrument fitted to it, and there was never enough undifferentiated consumer demand to force anyone to consolidate.

The result is a landscape organised by purpose rather than by proficiency. Once you can see the purpose classes, the design of any individual test stops looking arbitrary.

• Academic articulation — four skills, referenced to CEFR or ACTFL. CIMA, Hamza, NEWL, the Uzbek national certificate.
• Civil-service pay administration — receptive multiple choice, cheap, high-volume, objectively scored. YDS and e-YDS, and in part the Spanish EOI system.
• Defence and intelligence — referenced to the ILR scale, differentiated by dialect, and non-compensatory. The DLPT 5 plus a separate oral interview.
• Professional translation — directional translation under time pressure, marked as professional work. The CIOL Qualifications CertTrans and DipTrans.
• Workplace simulation — one integrated scenario rather than discrete items. France’s DCL Arabe.
• International civil service — a bespoke framework with eliminatory gates. The UN Language Proficiency Examination.

Five scales, and no common ruler

Five reference frameworks are in play, and they are not translations of one another.

CEFR, the Council of Europe’s A1–C2 scale, is the dominant reference for civilian Arabic testing. CIMA, Hamza, the Spanish EOI certificates and the Uzbek national certificate report CEFR levels natively — for those instruments the CEFR level is the score, not a conversion. ACTFL runs from Novice Low to Distinguished across eleven sublevels and is criterion-referenced, used in United States academia and industry. ILR is the United States federal scale, 0 through 5 with plus levels; the DLPT 5 reports only as far as 4. The United Nations grades its own examination at Level III of the United Nations Language Framework and does not map it to CEFR at all. And the UK’s RQF is not a proficiency scale in the first place: it locates a credential relative to a degree, which is why the DipTrans sits at Level 7 — master’s standard work — without certifying any particular CEFR band.

Concordance tables between these scales circulate widely and are useful for orientation. They are not jointly endorsed by ACTFL, the ILR and the Council of Europe, and the alignment is actively contested in the measurement literature. Turkish YDS letter bands are worse than contested: they are administrative pay tiers defined by statute, and were never designed to express CEFR levels at all.

The practical rule follows from that. Never use a cross-walk to satisfy a formal requirement. An institution that asks for a specific certificate wants that certificate, and a conversion table is not an argument you can win with.

What the last three years changed

Every Arabic instrument built in the last decade reports CEFR, and the two most ambitious of them arrived recently.

CIMA — the Certificat International de Maîtrise en Arabe, سمة in Arabic — came out of a 2015 partnership between the Institut du monde arabe and France’s national language-assessment body, piloted across seven countries in 2018, and launched publicly in April 2019. It certifies A1 to C2 in a single untiered test: two comprehension papers of forty-four items each, a writing paper of three tasks, and an eleven-minute individual oral. The certificate carries a global CEFR level, a per-skill CEFR level, and a percentage score on the two comprehension papers. It has been registered on France Compétences’ répertoire spécifique since 2021, and it runs through fifty-odd approved centres in around twenty countries, twice a year.

Hamza — اختبار همزة — is the clearest attempt yet to build an Arabic IELTS from inside the Arab world. Issued by the King Salman Global Academy for the Arabic Language and launched in December 2023, it covers CEFR A2 to C1 in seventy-five items across 155 minutes, and reports a scaled score from 130 to 200. It is delivered by computer at Saudi assessment centres, partner universities and by remote proctoring, and it has been administered in China, Indonesia and India as well as across Saudi and Gulf higher education.

The consequential development is not either launch on its own. It is that both are now associate members of ALTE, the European association of language testers — placing Arabic instruments inside the same professional standards body as Europe’s established examinations. Whether that consolidates the field or simply produces two rival CEFR-native brands, one Parisian and one Riyadh-based, is the live question in Arabic assessment right now.

The gaps in the map

Two things are worth saying about what a survey like this does not contain, because both are useful.

The first is Russia. There is no public standardized Arabic certification there comparable to CIMA or Hamza. Arabic is assessed inside degree programmes and internal attestation instead: MGIMO examines it through the state final examination at bachelor’s and master’s level, and the Diplomatic Academy of the Ministry of Foreign Affairs runs its own attestation for its own courses. The Russian teaching tradition in Arabic is real, philological and translation-centred, with vocalization and grammatical parsing at its core — but that is a pedagogy, not an examination regime, and it issues nothing portable.

The second is a handful of instruments most comparisons leave out. The King Salman academy runs a second examination alongside Hamza — an Arabic Competency Test introduced in 2022 and managed jointly with the Saudi evaluation commission. Attanal al-Arabi (التنال العربي) is an ALTE associate member offering a computer-adaptive test from A1 to C2 across five components, adding an explicit structure section to the usual four skills. And Avant Assessment’s Arabic Proficiency Test is used widely across United States schools for Seal of Biliteracy determinations, and belongs in any honest list.

The practical reading of both is the same. An absence in your jurisdiction is itself information: where no portable certificate exists, the receiving institution almost always runs its own internal examination, and the thing to ask for is that examination’s specification rather than a substitute certificate from somewhere else.

Volume follows statute, not pedagogy

Where testing is large, it is almost always because a law made it large.

Türkiye is the clearest case. The YDS is eighty multiple-choice items in 180 minutes covering vocabulary, grammar, translation and reading comprehension, with no negative marking. A single 2026 sitting drew more than 120,000 applicants, and Arabic was the second-most-selected language after English — ahead of German, French and Russian. That is the largest Arabic testing cohort of any instrument anywhere, and it exists because the civil-service foreign-language allowance is paid against the score under a decree-law. The mechanism is subtler than a one-off bonus: the result is valid for five years, and a holder who does not re-sit drops one level automatically, so a B-level holder becomes C-level and a C-level holder loses the payment entirely. The statute manufactures recurring demand rather than one-off demand.

Uzbekistan built the most aggressively incentivised Arabic market in the world by the same logic. A national multi-level certificate at 86% or above takes the maximum score for that subject in the state university entrance examination without sitting it, with proportional credit from 60%, and staff across state administration bodies holding a recognised certificate receive a salary supplement of up to twenty per cent.

The corollary is that these markets are policy-fragile. Uzbekistan revised its recognised-certificate list twice between 2024 and 2026 and removed benefits from two international certificates outright. Volume built on statutory incentive can disappear by decree, and a certificate that was worth a pay grade last year may be worth nothing this year.

Five instruments that are routinely misdescribed

Most of the bad advice in this field comes from a small number of specific errors, repeated. These five are worth having straight before you spend money or a year of preparation.

• The UK A Level in Arabic is not a four-skill qualification. Pearson’s qualification is formally titled Advanced GCE in Arabic (listening, reading and writing), and there is no speaking assessment at all — unlike A Level French, German or Spanish. Its two heaviest papers are translation in both directions plus written response to prescribed literary works. It is closer in construct to a translation diploma than to a proficiency test.
• The DLPT 5 does not test speaking. It assesses listening and reading only; speaking is measured separately through an approved oral proficiency interview. It reports ILR 0 through 4 — the scale runs to 5, but the test does not — and the structure is strictly non-compensatory, so strength in reading cannot offset weakness in listening.
• NEWL is a high-school instrument. It is endorsed by the College Board and accepted for credit or placement by more than sixty United States universities, but it is calibrated for the Novice High to Intermediate High range. It is not an advanced examination and should not be sold as one.
• YÖKDİL is not a substitute for YDS. Arabic YÖKDİL is offered only in the Social Sciences field, and it does not qualify for the Turkish civil-service language allowance — that is YDS territory exclusively. It satisfies institutional academic requirements and nothing beyond them.
• ACTFL ratings do not expire on their own. There is no intrinsic two-year window in the published materials; recency requirements are imposed by the receiving institution or employer and vary between them. Ask the receiver, not the internet.