How many learners reach basic fluency (B2) or near-native command (C2) — depending on where they start, using the CEFR classification.
The full CEFR fluency funnel charts a cohort of 100 complete beginners and watches them fall. About 55 make it out of A1. Twelve reach B1. Four reach B2. One makes C1. Roughly one in five hundred clears C2.
But most language learners aren't complete beginners. High school Spanish leaves you at A2. An intensive program drops you somewhere in B1. If you're already intermediate and wondering what the path to fluency looks like from here, this table conveys the challenge. The tables below renormalize the funnel at every entry point, so you can read your own odds directly.
The reason to put a number on it is that named obstacles are fixable ones. Knowing you're in the 22% who make it from A2 to B1 — or the 33% who clear B1 to B2 — reframes what's happening. This isn't a test of your aptitude. It's a map of the language obstacles every learner hits, each one with a structure and a route through it.
The underlying attrition figures — per-transition dropout rates and cumulative survivor counts — come from CEFR progression analyses consistent with Cambridge English's guided-learning-hour estimates. All figures are for Category I languages (Spanish, French, Italian, Portuguese, etc.). Category V languages (Mandarin, Japanese, Arabic, Korean) require roughly three to four times the hours; the shapes of the attrition curves are similar.
Table 1 — CEFR Attrition to B2 and C2, by Starting Level
Each row renormalizes its cohort to 100 people. Read it as: of 100 people who begin at this level and continue studying, how many eventually reach each destination?
| Starting Level | Real-world context | Of 100 starters → reach B2 | Of 100 starters → reach C2 |
|---|---|---|---|
| A1 | Complete beginner, first lesson | 4 | ~0.2 |
| A2 | High school / tourist-competent | ~7 | ~0.4 |
| B1 | Post-intensive or strong self-study | ~33 | ~2 |
| B2 | Basic fluency — already here | — | ~5 |
| C1 | Advanced / near-fluent | — | ~20 |
A few things worth sitting with: starting from A2 rather than A1 roughly doubles your odds of reaching B2 — not because A2 learners are better, but because the A1→A2 attrition (the Interest Obstacle and Habit Obstacle combined) has already been cleared. Starting from B1 makes a far more dramatic difference: one in three B1 learners eventually reaches B2, versus roughly one in fourteen A2 learners. The B1→B2 stretch is brutal, but if you're already standing at the bottom of it, you're already past the two biggest dropout events on the curve.
The C2 column is humbling at every starting point. Even C1 learners — people who are genuinely, demonstrably fluent — have only about a 1-in-5 shot at near-native command. C2 isn't a level you achieve by studying harder; it's mostly a matter of years of immersed living in the language.
Table 2 — Full CEFR Survival Matrix: Dropout Rates at Every Transition
This table answers a different question: of 100 people who start at level X, what percentage survive to each subsequent level? Read each row independently — each starting cohort is renormalized to 100.
| Starting Level → | Reach A2 | Reach B1 | Reach B2 | Reach C1 | Reach C2 |
|---|---|---|---|---|---|
| A1 (100 starters) | 55% | 12% | 4% | ~1% | ~0.2% |
| A2 (renorm to 100) | — | ~22% | ~7% | ~2% | ~0.4% |
| B1 (renorm to 100) | — | — | ~33% | ~8% | ~2% |
| B2 (renorm to 100) | — | — | — | 25% | ~5% |
| C1 (renorm to 100) | — | — | — | — | ~20% |
The diagonal of this table is the probability of clearing the next obstacle from each starting point: 55%, 22%, 33%, 25%, 20%. The B1→B2 transition is the odd one out — it's actually easier in relative terms than A2→B1, even though it demands more total hours. The reason is selection: the learners who reach B1 are not a random sample of people who once downloaded an app. They're the ones who already survived the two biggest dropout events on the curve. They're more motivated, more self-directed, and more likely to have built the habits that get you through a long plateau. The funnel self-selects for exactly the trait the Big Obstacle requires.
The A2→B1 row is the grimmest in the matrix. Of 100 people who finish high school Spanish — three or four years of structured instruction — only about 22 will ever reach the level where they can hold a real conversation. The rest plateau at "good enough for a restaurant" and never push further. That's not a failure of those learners; it's what happens when a learning environment provides no middle-ground content and no mechanism for forced output, then disappears entirely.
How to Read These CEFR Attrition Estimates
The figures above are approximations derived from CEFR-progression research and consistent with Cambridge English guided-learning-hour data. Exact dropout rates vary by language, study population, instruction context, and how strictly "reaching a level" is defined. What doesn't vary much is the shape: steep early attrition, a second cliff at A2→B1, relative improvement in survival odds once learners clear B1, and a near-impassable wall between C1 and C2 without immersion.
Use these numbers as a map, not a guarantee — but use them. Knowing you're in the top 7% of A2 starters the moment you reach B2 is the kind of fact that reframes a plateau. You're not failing a test. You're on a path that most people left years ago.
What the Numbers Don't Capture
Survival rates are population statistics. They describe what happens when a learner hits a wall without a name, without a tool that fits, without any way to tell whether a plateau is permanent or just the B1→B2 grind everyone goes through. The learners counted in these tables mostly quit not because they lacked talent, but because they ran out of map.
The attrition numbers are real, but they aren't fixed. The B1→B2 climb — the one where one in three learners eventually makes it — is hard partly because it's invisible: you put in an hour and have nothing concrete to show for it. Atlas Runa is built for exactly that stretch. The Progress Log turns a week of uneven practice into visible evidence, so you stop measuring progress by whether something felt productive and start measuring it by what actually happened. The Reader keeps surfacing level-matched content so the content gap doesn't stall you. Vocabulary review runs quietly in the background so the words you're about to forget come back before they're gone.
The CEFR funnel doesn't have to be your funnel. The language obstacles that cause most of this attrition have names and routes through them — and knowing which one you're on is most of the battle.
The obstacle-by-obstacle breakdown of why each attrition event happens — what the research says, what it feels like from the inside, and what actually moves the needle — lives in the CEFR fluency funnel.