HSK Speaking Level Descriptions

Compare the three HSK Speaking learner profiles, then separate vocabulary familiarity from speaking.

Three-step decision diagram: read the learner profile, collect speaking evidence, then confirm unresolved official questions.
Editorial synthesis, not an official diagram or validated placement test.
On this page
  1. What the three learner profiles actually describe
  2. Keep three kinds of evidence separate
  3. Three profiles, one question each
  4. A score would hide what the sheet is for
  5. Turn the worksheet into a next step
  6. Source date and limits

Chinese Testing International’s HSK Speaking overview describes an assessment of oral Chinese with elementary, intermediate and advanced levels, conducted through audio recording. That makes speaking evidence relevant to your preparation decision. The worksheet below is editorial guidance, not an official or validated placement instrument, and cannot predict a pass.

What the three learner profiles actually describe

The first three columns summarize the publisher’s captured descriptions. The final column is our interpretation: a question to investigate, not an additional official requirement. Keep the approximate numbers attached to their context rather than turning them into precise cutoffs.

Official level pageDescribed learning historyApproximate vocabulary masteredEditorial question to investigate
ElementaryOne to two semesters, with 2–3 lessons per weekAround 200 commonly used wordsWhat can I express aloud with the words I recognize?
IntermediateOne to two academic years, with 2–3 lessons per weekAround 900 commonly used wordsWhere does my speaking evidence support—or fail to support—my vocabulary estimate?
AdvancedMore than two academic years, with 2–3 lessons per weekAround 3,000 commonly used wordsWhich parts of my apparent readiness remain assumptions?

The publisher frames these as intended candidate profiles. The supplied excerpts do not define how vocabulary mastery is measured or how long a lesson lasts. Consequently, neither an app’s recognized-word total nor a calendar of study can be treated as an exact conversion into one of these levels.

Keep three kinds of evidence separate

Create a short decision sheet with three sections. Do not combine them into a numerical score: the purpose is to expose missing evidence, not manufacture a placement result.

1. Vocabulary familiarity

  • Record where your estimate came from: a course list, flashcards or another personal record.
  • Distinguish words you recognize from words you can retrieve while speaking.
  • Mark uncertainty instead of rounding your estimate up to match a profile.

This section describes your estimate and its limitations. It does not establish that you have met the publisher’s meaning of vocabulary mastery.

2. Learner-created speaking samples

  • Record a short response to a familiar topic, such as introducing your work or describing a recent activity.
  • Keep the prompt and note whether you used a script, notes or repeated attempts.
  • Listen for observable issues: a missing word, an abandoned sentence or a point you could not explain.
  • Write a specific practice action beside each issue, rather than assigning yourself an official level.

These are editorial practice suggestions, not captured examination tasks or scoring criteria. A sample is useful here because it preserves something you actually said under stated conditions. It is not equivalent to an assessed test performance.

3. Questions requiring official confirmation

  • Which current official task descriptions and assessment criteria apply to the level you are considering?
  • Is there current official guidance that clarifies how to interpret the learner profile?
  • Does any organization requesting a result specify a particular level or requirement?

Keep unanswered questions visible. The supplied profile excerpts do not contain detailed speaking-task instructions or a scoring rubric. Do not fill those gaps with assumptions from vocabulary totals.

Three profiles, one question each

The table’s last column is the point of the whole article: each profile ends with a question to investigate, not a verdict. The elementary profile asks what you can express aloud with the words you recognize; the intermediate profile asks where your speaking evidence supports — or fails to support — your vocabulary estimate; the advanced profile asks which parts of your apparent readiness remain assumptions. The profiles describe learning history and approximate vocabulary together — semesters and lessons per week beside the totals — and the article insists the two stay attached to their context rather than becoming precise cutoffs. A profile is a candidate to investigate, and the question beside it is the investigation.

A score would hide what the sheet is for

The decision sheet has three sections — vocabulary familiarity, learner-created speaking samples, and questions requiring official confirmation — and the article says not to combine them into a numerical score. A score would hide exactly what the sheet is for: exposing missing evidence. Vocabulary familiarity asks where your estimate came from and whether you can retrieve words while speaking, not just recognize them. Speaking samples preserve something you actually said under stated conditions, with the prompt, the notes and the repeated attempts recorded beside it. Unanswered questions stay visible because the supplied excerpts contain no scoring rubric. Three sections, kept separate, so the gaps stay in view — and the hypothetical example of two learners with the same word total shows why: the totals match, but the samples document different conditions, and only the samples say what each learner can actually do.

Turn the worksheet into a next step

Your evidence situationEditorial next step
Only a vocabulary estimateCollect a speaking sample before drawing a readiness conclusion.
Vocabulary and speaking evidence point in different directionsDocument the mismatch and choose a concrete practice target; do not force a level label.
Your records resemble a learner profileTreat that level as a candidate to investigate against current official guidance, not a confirmed placement.
A consequential requirement remains unclearSeek confirmation from the relevant official source or requesting organization before deciding.

Hypothetical example: Two learners each report recognizing roughly 900 words. One records a response without notes; the other reads a prepared script. Their totals match, but their samples document different conditions. Neither record alone proves intermediate placement or predicts a pass. The useful next step is to clarify what each learner can produce independently and what official guidance still needs checking.

Source date and limits

The supplied source-check date is October 7, 2026; the cited excerpts were captured that day in UTC. This is an interpretation of captured documents, not a live service or verified availability check. It does not address schedules, booking or result-validity periods. Refresh the article if the level names, learner profiles, vocabulary figures or recording description change. The evidence expires on October 8, 2026, at the source-specific times recorded with the captures; this draft does not extend that expiry.

Related reading: HSK & Chinese Language Exams: Preparation Hub · Build a Chinese Communication Task Map with New HSK's Five Components