2026: The End of Elementary School Tutoring

The era of excessive tutoring for elementary school students can finally end in 2026. In the age of AI, extracurricular tutoring—aside from babysitting—serves almost no purpose.

Over the Dragon Boat Festival, I cross-analyzed 3,978 publicly available exam papers from Chinese elementary schools (grades 1–3). My finding: “studying hard” as we know it is just not that effective.

If you drop tens of thousands of yuan each year on tutoring, ask yourself honestly: are you really getting a “tutor” — or are you buying a self-soothing “I’ve done my part” sticker?

1. The number that surprised me

I went through every single Chinese and math paper I could find for grades 1–2, scanning them one by one for statistics.

In the end, I analyzed 3,536 papers that had extractable text out of the original 3,978. Among them, 1,604 fell under the “end-of-term category” — i.e., real tests, final exams, and mock exams.

I initially wanted to find some “pattern of difficult problems.” But after scanning, the most direct takeaway was:

The high-frequency test points in grades 1–2 final exams are mostly not “beyond-the-textbook” problems, but rather contextualized, situational, and integrated versions of the basic content in the textbook.

In plain English: No matter how you roll it, they’re still testing what’s in the textbook — just in a different format or scenario.

That conclusion itself isn’t shocking. What is shocking is how wildly it contradicts the level of “intense competition” I see every day.

2. What the data looks like

Let me lay out the most boring part first, because my argument builds on it.

For Chinese, regardless of grade or semester, the top recurring items are always these: pinyin, vocabulary, sentences, characters, and reading comprehension.

• In second grade (second semester), “vocabulary: word grouping, synonyms/antonyms, collocations, measure words, categorization” appeared in 0.97 of the papers — 266 out of 275 valid papers tested it.
• Sentences: 0.92; reading comprehension: 0.91; pinyin: 0.89.

What does that mean? If you randomly pick a second-grade final paper, you are almost guaranteed to see word grouping, synonyms/antonyms, measure words — the kind of stuff drilled endlessly in the textbook’s language garden. No mystery, no obscure topics.

Math is even more straightforward. For first and second grade finals, the dominant topics are: word problems, computation, geometry/shapes, and number sense.

• First grade (first semester): word problems appear in 0.94 of papers, computation in 0.82.
• Second grade (second semester): computation and word problems both hit 0.94.

What about “advanced” content? I specifically checked how often “multiplication/division within tables / division with remainders” appeared in first-grade papers:

• First grade (first semester): only 8 out of 119 valid papers had it—frequency labeled as “occasional.”
• First grade (second semester): 5 out of 130 papers.

In other words, genuinely advanced content is extremely rare and marginal in first-grade final exams.

In my report I singled out a few examples: parallel/perpendicular, net diagrams, counting triangles, row/column coordinates—these are not the main focus of finals. Most come from thinking assessment papers, competition-style drills, or content that leaks in from higher grades. They should not be treated as the battleground for first-grade tutoring.

3. So what is the “intense competition” really about?

Now let’s put two things side by side:

On one hand, 1,604 exam papers tell me: they test the basics. Consistently, reliably, without surprises.

On the other hand, in reality there’s an overwhelming push for: learning ahead, solving hard problems, enrolling in advanced classes, racing into Olympiad thinking.

The first thing I couldn’t figure out is: A huge amount of this “intense competition” is aimed at things that aren’t even on the exam.

Kids are pushed to learn net diagrams, count triangles, and use row/column coordinates when the final exam barely has any such questions. Anxious parents send their kids to advanced classes that target content only “occasionally” seen in the current grade. This isn’t effort—it’s wasting energy on things that won’t yield scores, and it often ruins the child’s appetite for learning in the process.

If “intense competition” is for exams, it’s aiming at the wrong target. If it’s not for exams, then what’s the point? That’s a bigger topic—mostly it’s about easing the parents’ anxiety, not building the child’s ability.

4. About “focusing only on school grades”

Someone might argue: “Fine, since exams test only textbooks, I’ll just grind the textbook and grind the score. No harm in that, right?”

This is exactly the second thing I want to question.

Look at the data structure: the test points are so narrow, coverage rates are above 0.9. What does that mean? It means the gap between a perfect score and 95 points is not about ability—it’s about proficiency and attention to detail.

A second-grader knows word grouping, synonyms/antonyms, measure words. The few points they lose are probably because of a missing punctuation mark, a copied wrong character, or rushing through a reading question. This is a normal developmental fluctuation, not a knowledge gap.

But the “focus-only-on-grades” logic does this: to lock in those floating 5 points, they throw in double the time, endlessly drilling the same types of problems the child already knows.

The absurdity is: You spend your most valuable time to buy the least valuable 5 points.

Those 5 points seem important on a first-grade final paper, but in the context of a child’s growth over a decade, they mean almost nothing. The time squeezed out by those 5 points—reading for fun, daydreaming, tinkering with things they love, completing a project from start to finish—that’s what’s truly scarce and irreversible.

The narrower the exam points, the lower the cost-effectiveness of obsessing over grades. Because you’re not competing for “can you do it?” but for “are you fast enough and steady enough?” Spending large chunks of childhood on something with diminishing marginal returns—I can’t justify its necessity.

5. In the AI era, the foundation of the tutoring industry is collapsing

So far I’ve only talked about “competing in the wrong direction.” What really makes me believe tutoring can end is AI.

Tutoring historically depended on two asymmetries: information and human labor. The teacher had methods you didn’t, the patience to explain, and the time to supervise. What parents paid for was essentially: someone who knows more and has more time to feed the knowledge to my child.

But by 2026, AI already covers all of that, and more:

• Explaining? A second-grade word problem they don’t get? Snap a photo, toss it to AI. It can explain it in three ways until the child understands—no annoyance, no impatient face, available at midnight.

• Generating and grading drills? All those basic test points with coverage rates above 0.9? AI can generate endless variations, grade them instantly, and point out if the mistake is in reading the problem or writing the answer.

• Personalized instruction? A tutoring class has one teacher for 20 kids—pacing is a compromise. AI is truly one-on-one; it stays longer where the child is stuck.

Look at these pieces: the money tutoring centers used to charge is now essentially replaced by AI—cheaper, more patient, available on demand.

So when parents spend tens of thousands sending their child to a tutoring center, what are they actually buying?

At its core, they’re buying “someone to watch my kid for a few hours” —a safe, supervised, and “looks like learning” babysitting service. That’s a real need—I won’t deny it. But call it babysitting, not education. It solves the problem of parents being busy, not the problem of the child’s abilities.

6. So how should we use this data?

I’m not saying “basics don’t matter.” On the contrary, the data’s clearest fact is: basics are the backbone. The textbook is the backbone.

So what we should actually do is disappointingly simple:

Cover the textbook basics—pinyin, characters/words, sentences, computation, word problems—thoroughly. Make sure the child understands, can do them correctly, and doesn’t trip over reading or writing. That’s enough to secure an A in grades 1–2.

Then, stop wasting energy on learning ahead or advanced drills. The final exam barely tests that stuff, and the child is young—let them progress step by step. Save the time for things that can’t be measured by a test paper but are more valuable.

When I scanned these 1,604 papers, I was hoping to find a shortcut to squeeze out a few more points for my kid. What I found instead was a reason to let go:

Exams are simpler than we think, and we are more anxious than our children need us to be.

7. Rebuilding learning with gamification: getting kids to “play” their way to knowledge

I’ve talked a lot about “what not to do.” So what should we do?

My answer is the product I’m building.

First, a question: Why are games so addictive? Because they hit two innate human desires: curiosity and a sense of achievement. Unlocking levels one by one, instant feedback, difficulty set just at the “within reach” zone—that mechanism makes people engage actively, without any external pressure.

The core purpose of my product is to fully integrate this addictive game logic into children’s learning scenarios.

Note: this is not about cramming knowledge into a game shell—I’ve seen too many “rebranded drills.” I want children to naturally achieve their learning goals within an immersive, edutainment experience:

• What needs to be mastered are those basic textbook exam points with coverage rates above 0.9—pinyin, vocabulary, computation, word problems—all included.
• The method of mastery is play, not grind.

Because, compared to forced instruction, getting a child to actively immerse themselves is the most efficient way to learn.

Tutoring can end in 2026. But learning won’t end—it just needs a new format that children are willing to keep playing.

Inside this app, I’ve also embedded a “game” logic to make “studying” addictive for kids.