How we work it out
Every number the calculator shows is one published figure and four coefficients. The published figure is sourced. The coefficients are our estimates, and this page says so for each one rather than dressing them up as research.
1. The baseline — published, not ours
The US Foreign Service Institute publishes how many class hours its students need to reach ILR Speaking-3 / Reading-3, described as General Professional Proficiency. That is the only number here we did not invent.
| Group | Four-category | Five-category | Hours |
|---|---|---|---|
| similar | Category I | Category I | 600 |
| moderate | Category II | Category III | 900 |
| hard | Category III | Category IV | 1100 |
| super hard | Category IV | Category V | 2200 |
Two numbering schemes circulate for this same data — the State Department currently publishes four categories, most community sources use five. The hours are identical, so we key on hours and show both labels rather than picking one and calling it official.
2. Which CEFR level you actually want — estimated
FSI measures one endpoint. Most people are not aiming at it. We scale the baseline using the shape of the published Cambridge guided-learning-hour bands, normalised against C1.
- A1×0.13
- A2×0.27
- B1×0.5
- B2×0.73
- C1×1
- C2×1.47
Estimated. Cambridge guided-learning-hour bands per CEFR level, normalised against the C1 band. Captures the non-linear shape of the curve; absolute hours come from FSI, not from Cambridge. Checked against the published table on 2026-08-08: every factor above lands inside its band — A2 202 against 180-200, B1 375 against 350-400, B2 548 against 500-600, C1 750 against 700-800, C2 1102 against 1000-1200.
This rests on treating FSI's endpoint as roughly C1. No official ILR↔CEFR crosswalk exists. Published academic tables place ILR 3/3+ at C1; others place C1 nearer ILR 4.
3. How you study — estimated, and level-dependent
A single multiplier per method misprices both ends. An app's weakness is that it never makes you produce language under pressure and never corrects you — which barely matters at A1 and matters enormously at B2. Immersion runs the other way.
| Method | At A1 | At C1 | Ceiling |
|---|---|---|---|
| Intensive class (FSI-style) | ×1 | ×1 | C2 |
| 1-on-1 tutor | ×1.05 | ×1.05 | C2 |
| Structured self-study course | ×1.25 | ×1.7 | C1 |
| App only | ×1.3 | ×3.5 | B1 |
| Living in-country | ×0.9 | ×0.75 | C2 |
Estimated. Reasoned from what each method provides — corrective feedback, forced output, unpredictable input — relative to the FSI classroom the baselines were measured in. No published research assigns coefficients to these categories.
The ceiling column matters more than the multipliers. When your target sits above your method's ceiling, the calculator says so instead of quoting a very large number of hours. Apps are efficient for early vocabulary and basic patterns, and get steadily worse from there: they are recognition-heavy, so you pick the right answer far more often than you produce language under pressure. Genuinely hard to push past B1 with alone.
4. Languages you already speak — estimated
A Spanish speaker learning Portuguese is not doing the same work as a monolingual English speaker. Transfer effects in third-language acquisition are well attested; their size is not published as a coefficient anyone can cite, so ours are set conservatively and capped so that stacking several known languages cannot produce an implausible discount.
Estimated. Transfer effects in third-language acquisition are well attested; no published coefficient exists. Values are set conservatively — shared branch gives the largest reduction, shared script a smaller one, and the total is capped so stacking several known languages cannot produce an implausible result.
5. HSK, for Chinese
Hanban's claimed 1:1 HSK↔CEFR equivalence was rejected by CEFR bodies. German and French associations of Chinese-language teachers place old HSK 6 around B2–C1, not C2. No official mapping exists.
We map low. Telling you HSK 6 is C2 would shorten every estimate on this site, which is precisely why we do not.
What this cannot tell you
Nothing here predicts your outcome. It is a planning figure built from a published average and four judgement calls, and individual variation swamps all of it. Treat a result as an order of magnitude, not a date.
We also do not test these courses at scale, which is why you will not find star ratings anywhere on this site.
Sources
Everything the model rests on. Two of these are published figures; the rest document where no official answer exists, which is why several of our coefficients are labelled estimates.
- Foreign Language Training — US Department of State, Foreign Service Institute
The class-hour bands every estimate on this site starts from, and the definition of the proficiency they lead to (ILR Speaking-3 / Reading-3).
- FSI Language Difficulty Ranking — fsi-language-courses.org
Which languages sit in which band. The State Department publishes the categories as an image, so per-language assignment comes from independent reproductions that agree with each other and with the official hour bands.
- Guided learning hours — Cambridge English
The shape of the curve between CEFR levels. Cambridge publishes cumulative hours per level (A2 180–200, B1 350–400, B2 500–600, C1 700–800, C2 1,000–1,200); our level factors are normalised against the C1 band and every one lands inside its published range.
- ILR scale — Wikipedia
That no official ILR-to-CEFR crosswalk exists. Published academic tables place ILR 3/3+ at C1; others put C1 nearer ILR 4. We take the lower reading, which makes our estimates the more conservative of the two.
- Zhongwen Shuiping Kaoshi (HSK) — Wikipedia
That Hanban's claimed one-to-one HSK-to-CEFR equivalence was rejected by CEFR bodies and is disputed by the German and French associations of Chinese-language teachers, who place old HSK 6 nearer B2–C1 than C2.
Written and maintained by Kevin Fan. Found an error, or a source we should be using? Try the calculator and tell us what looks wrong.
https://fluencyclock.com/how-we-test