Hiring · July 30, 2026 · 10 min read

TOPIK levels in hiring: what a Level 6 really proves

TOPIK in hiring is a strong signal with a blind spot: the main test never measures speaking, so a Level 6 holder may never have had their spoken Korean tested.

← Part of The five pillars of hiring: what assessments measure

If you hire foreign talent into Korea, or screen non-native Korean speakers for roles at a Korean company, the TOPIK line on a CV is probably carrying more weight in your decision than you have examined. It deserves respect. A high TOPIK level is a real, hard-won achievement, and a candidate who holds one has proven something genuine about their Korean. But using TOPIK in hiring well means understanding exactly what the certificate proves and what it structurally cannot. The stakes are rising: as Korean firms compete harder for international staff and language expectations tighten across the market, an over-read or under-read certificate becomes an expensive habit. This is the employer's mirror of the argument we made about JLPT in hiring for Japan — same discipline, a different test with its own blind spot.

This is not an argument against TOPIK. It is an argument for reading it accurately. The certificate has two limits that no amount of respect changes: the main test never measures speaking, and a result is valid for only two years yet routinely appears on CVs long after. Once you see both clearly, the right use becomes obvious — treat TOPIK as a signal to filter people in, then assess the actual working Korean the role runs on. That is the honest position, and it is fairer to candidates than either dismissing the certificate or trusting it blindly.

What do TOPIK levels actually certify?

TOPIK, the Test of Proficiency in Korean, is administered by NIIED, Korea's National Institute for International Education. It reports six levels across two separate tests. TOPIK I covers levels 1 and 2, the beginner band, and tests listening and reading only. TOPIK II covers levels 3 through 6 and adds a writing section, where level 6 is the highest grade and signals comprehension of Korean in professional, academic and abstract settings. Your level is not chosen in advance; it is assigned by total score against thresholds NIIED publishes. A level, then, tells you a candidate cleared a band — a credible, standardised achievement worth having near the top of a funnel.

One fact about the exam's design matters more than any other for hiring, and it is easy to miss. Neither TOPIK I nor TOPIK II tests speaking. Levels 1 and 2 measure listening and reading; levels 3 to 6 add writing. None of the six measure spoken Korean at all. So the main TOPIK certifies comprehension and, at the higher levels, written production — but it is silent on whether a candidate can actually hold a conversation, present in a meeting, or negotiate on a call. That silence is structural. It is not a flaw in the exam; it is simply the boundary of what the main test was built to measure.

None of this diminishes the achievement. Reaching level 6 takes years of disciplined study, and it is a reliable sign of serious commitment as well as genuine comprehension and written ability. The mistake is not respecting the certificate; it is quietly treating a high level as proof of a rounded working fluency the main test was never designed to certify. TOPIK is precise about what it covers. Hiring goes wrong when we are vaguer about it than the exam itself is.

Why can a Level 6 holder have untested spoken Korean?

Because the spoken test is a separate exam, and it is still scaling. NIIED piloted a dedicated TOPIK Speaking test in 2023, computer-based and reported separately from the main grade. It is being rolled out gradually, with capacity growing toward tens of thousands of seats over the coming years — the publicly stated target is around 50,000 seats by 2027. But it is not yet the default, and it is not the test most candidates sat to earn the level on their CV. So a candidate can hold TOPIK level 6 and have never had their spoken Korean formally assessed by anyone.

This is the trap for employers. The level number is high, the paper is impressive, and the natural assumption is that someone who reads and writes advanced Korean can also speak it fluently. Often they can. But the certificate does not tell you that, and comprehension and production are different muscles. A candidate can decode a dense report and still stumble in a live discussion, because the exam never asked them to speak. Do not read the TOPIK Speaking pilot as a requirement, either — as of now, you should not assume employers or visa routes broadly require it. It exists, it is growing, and it is worth knowing about. It does not yet fill the gap for you at the point of hire.

Why does the two-year validity matter on a CV?

A TOPIK result is valid for two years from its announcement date, per NIIED. TOPIK is used widely for university admission, some visa categories and permanent-residency applications, and that two-year window is built for those uses. Hiring is different. CVs routinely carry TOPIK grades earned well beyond two years ago, and a level 5 from years back reads on paper exactly like a level 5 from last month. The certificate freezes a single exam window; the candidate in front of you is not that window.

This matters because proficiency is not fixed. Research on language attrition — the kind published in journals such as Language Learning — finds that productive skills, speaking and writing, tend to fade before receptive ones when a language goes unused, and that higher attainment offers some protection against that fade. We quote no percentages, because the honest figures vary by person and context. The direction is what counts: someone who passed a high level, then worked for years in an English-only environment, may sit below their certificate today, especially on the production side. Someone at a lower level who has lived and worked in Korean since may be far stronger than their paper suggests. When the exam date is old, the level quietly overstates its own relevance.

What does 'work language' actually mean, role by role?

Work language is the specific bundle of reading, listening, writing and speaking a particular role runs on — and the mix differs sharply by job. That is why a single TOPIK number is a poor proxy for readiness across roles. The useful question is not 'what level?' but 'what does this person have to do in Korean, all day, and can they do it?' Three cases show how differently the answer lands.

A support agent lives on production under time pressure. They read a customer's message, understand it, then write a reply that is correct, clear and pitched at the right level of politeness — honorifics carry real weight in Korean, and the wrong register lands badly — often dozens of times a day. TOPIK II covers the writing here, and covers it fairly well. This is why hiring a customer support representative should start from the language the role produces, not merely the one it understands. But the moment that role also takes phone calls, the certificate goes quiet, because the calls are spoken and the main test never was.

An engineer's mix is different again. They may read Korean specs and chat threads all day — heavy comprehension — while producing mostly in code and short written updates, with speaking concentrated in stand-ups and design reviews. The reading bar can be high while the spoken-polish bar is moderate. A sales rep inverts it entirely: their value can rest almost wholly on spoken persuasion and the trust built in a room, exactly the skill TOPIK does not touch. A blanket 'level 5 required' treats these three as one problem. They are not. The certificate measures the skills all three happen to share and stays silent on the ones that separate a strong hire from a weak one in each case — most sharply, speaking.

How do you assess the working Korean the certificate misses?

Put the candidate in front of the work. The most predictive way to test production is a job-shaped task. Ask a support candidate to write the reply to a real-looking customer complaint, in the register the role demands. Ask an engineer to summarise a short Korean spec in writing. For any role that lives on conversation, run a short spoken task — a mock call, a two-minute answer to a client-style question — because that is the one thing the main TOPIK could never show you. Each of these observes the actual skill under conditions close enough to the job that the result predicts performance rather than approximating it.

This is the logic of work-sample tests applied to language: the closer an assessment sits to the real task, the better it predicts. A comprehension-and-writing certificate is a proxy; an observed spoken answer, scored against a consistent rubric, is evidence — and it is the evidence TOPIK structurally cannot give you. The two are complementary. Use the certificate to filter in, since it is cheap and standardised, then spend your assessment effort on the skills that decide whether someone can do the job, with speaking first in line for anything conversational.

This is where a structured language gate earns its place. Rather than reading Korean ability off a CV line, you can set a hard language requirement at capability level — reading, listening, writing and, crucially, speaking where the role needs it — as a prerequisite a candidate clears before the role-specific questions begin. Our note on language proficiency in hiring sets out the wider case: language, where the role depends on it, is a job skill to be measured directly, not a claim to be trusted. Assessing it natively — including spoken and audio modalities a written multiple-choice item cannot reach — is the whole point, and you can see the approach in a sample assessment.

How do you set a fair TOPIK bar without over-filtering?

By matching the certificate requirement to what the role genuinely reads, writes and hears — and no higher. Demanding level 6 where level 4 plus strong verified speaking would serve the work is not rigour; it is over-filtering. It shrinks an already tight pool of Korean-capable international talent, and it biases systematically against capable people whose certificate is out of validity or understates their current ability. Any specific level you have seen quoted as an employer standard is a commonly cited convention, not an official rule — NIIED sets the levels, not hiring bars. The bar should track the job, applied identically to everyone who reaches it.

The fairness problem cuts both ways. Set the bar too high and you screen out strong candidates for a level the work never needed. Read the certificate too generously — especially on speaking, which it never tested — and you advance people whose spoken Korean nobody checked. The discipline that resolves both is the same one behind skills-based hiring: decide what the role actually requires, measure that directly, and keep proxies in their place. A TOPIK level, used as a filter-in signal and paired with a real production assessment, is a good proxy. Used as the whole decision, it quietly does unfair work — sometimes too strict, sometimes too lenient, always resting on an exam that may be out of validity and never heard the candidate speak.

Getting this right also protects the candidate experience, which matters more when the talent pool is competitive. A candidate asked to demonstrate the actual work — write a real reply, answer a real spoken prompt — generally finds it fairer than being filtered on a certificate alone, because it lets them show what they can do rather than only what they once passed. And this is not Korea-specific reasoning. The same filter-then-verify logic runs through the whole family of language certificates in hiring, from JLPT for Japan to the frameworks behind them; a certificate names a level, and the work settles what the level cannot.

A TOPIK level proves a candidate read, listened and — from level 3 up — wrote Korean on one exam day, perhaps outside its two-year validity, and says nothing about whether they can speak it in the meeting your job runs on. Respect the signal, then test the work.

If you want to see how job-shaped language assessment works in practice — a written reply or a spoken answer scored against a consistent rubric rather than inferred from a certificate — you can try a sample assessment or book a demo before you change anything about how you screen for Korean.