Khan Academy, IXL, or a Live Tutor: Which One Actually Closes the Learning Gap?

Your child is behind. Maybe you've known it for a while, even if the report card was the first time it was official. So you did what most parents do: you typed something into Google, found Khan Academy and IXL, and wondered if one of them might be enough.

Both are legitimate. Both have research behind them. Both are used by students in schools across the country. And if your child has a mild gap or just needs extra practice on a skill they mostly understand, either one might genuinely help.

But "closing a learning gap" is a specific problem. It's not the same as reviewing material you already know, or getting extra practice when you're mostly on track. A real gap — the kind where a 6th grader is working at a 4th grade level, or where a child hits a wall every time fractions appear — requires something different. This article is about what each option actually delivers, where each one hits its limit, and what the research says about which approach tends to move the needle.

According to NWEA's recent assessment data, the average student still needs the equivalent of nearly 5 additional months of reading instruction and 4.4 months of math instruction to return to pre-pandemic levels. And those are averages. For students who were already behind, the gap is wider.

NWEA's most recent dashboard, updated in early 2026, found that reading achievement remains stalled with little rebounding to pre-COVID levels. That context matters when you're deciding whether a practice platform is going to be enough.

Khan Academy is genuinely impressive for what it is: a free, self-paced library of instructional videos and mastery-based exercises covering K–12 math, science, history, and more. For a student who mostly gets a concept but needs another explanation — a different angle, a slower walkthrough — Khan is often exactly right. Homeschool families and classroom teachers use it as a reliable supplement for this reason.

The platform has real research behind it. A randomized-controlled trial involving nearly 11,000 students in grades 3–8 found that students who used Khan Academy as part of a mastery learning intervention improved their end-of-year math scores by between 0.12 and 0.22 standard deviations compared to a control group. That's a meaningful, measurable result.

But there's an important footnote in Khan's own research: most students only reach about 86% of their target growth in a given year, and the platform explicitly notes that significant improvement requires consistent effort. That "consistent effort" piece is where the model shows its limits.

Khan is built for student-directed use. The platform teaches and assesses, but it doesn't monitor whether your child is actually engaging rather than half-watching the video. For students who are already behind and struggling to maintain motivation, self-direction is exactly the skill they often don't have yet. The model puts the responsibility to learn on the student — and for students who struggle to work independently, that structure doesn't hold.

The deeper issue is diagnostic. Khan can tell you that your child got 7 out of 10 questions wrong on a unit about fractions. It can't tell you why. It can't distinguish between a child who misunderstands numerators, a child who forgot how to find a common denominator, and a child who never really understood what a fraction means in the first place. Without that diagnosis, the platform serves up more practice. But practice on the wrong foundation doesn't build a stronger one.

IXL is used by roughly one in four U.S. students and is built around adaptive practice — a SmartScore algorithm that adjusts difficulty in real time based on how the student is performing. It covers math and English language arts from pre-K through 12th grade, and its diagnostic tool correlates highly with standardized assessments like NWEA MAP.

The research on IXL is substantive. A three-year study in Oklahoma evaluated by SRI International found that students in schools using IXL consistently outperformed comparable schools that did not use it. A 2024 home-use study found that students who reached proficiency in at least two IXL skills per week significantly outperformed those with minimal usage.

Compared to Khan, IXL is more structured and more responsive. The adaptive algorithm means your child isn't just working through a static playlist — the platform is (to a degree) adjusting to them. For building fluency and reinforcing skills the student has already been introduced to, it's a strong tool.

IXL Math is designed as interactive practice on singular skills — it does not blend multiple skills together in the way real-world problems and test questions do. The SmartScore system penalizes wrong answers in a way that some students find discouraging, and the platform has no mechanism to explain a concept differently when a student is stuck. It can sense that a student keeps getting a question type wrong; it can't figure out what's causing it.

For students with genuine foundational gaps — not just gaps in practice volume, but in conceptual understanding — adaptive practice on top of a shaky foundation produces more frustration than progress. The platform assumes the student understands what the skill is; it's built to refine and reinforce, not to reteach from scratch.

This is where the comparison gets sharper. Studies that pool the results of dozens of tutoring trials consistently find that students who receive live tutoring make meaningfully more progress than those who don't — roughly the equivalent of several additional months of learning in a single school year. That effect is two to three times larger than what the best research on platforms like Khan or IXL typically reports, making tutoring one of the most effective interventions identified in education research literature.

The defining characteristic of high-impact tutoring is real-time responsiveness. A live tutor can monitor understanding closely, provide immediate feedback, and adjust instruction in response to student progress — qualities that make tutoring particularly well suited to addressing unfinished learning and gaps in foundational skills.

EdResearch for Action, reviewing the design principles for high-impact tutoring, found that the most effective programs are characterized by strong tutor-student relationships, frequent sessions, and data-informed practices — and that virtual tutoring by a live tutor can be just as effective as in-person models when well-structured.

One thing that consistently shows up in the research: ensuring a student has a consistent tutor over time facilitates better outcomes, because that relationship allows the tutor to build a deeper understanding of exactly where that specific student's learning breaks down.

This isn't a case where one option is universally better. It depends on what's actually causing the gap.

Khan Academy or IXL might be enough if:
• Your child understands the concept but needs more practice reps to build fluency
• The gap is recent (a unit or two where something didn't stick) rather than cumulative
• Your child is self-motivated and will actually use the platform consistently
• You're looking for low-cost reinforcement alongside other support they're already getting

A live tutor is probably necessary if:
• Your child has been behind for more than one school year
• They can't explain a concept even after watching multiple explanations
• Their confidence around the subject has started to break down (giving up quickly, refusing to try)
• The gap involves a foundational skill that everything else is built on (like place value, or decoding in reading)
• You've tried apps or platform-based practice before and seen little change

The honest version of this: apps and adaptive platforms are tools for maintaining and reinforcing learning. They're not well-designed for diagnosis, for reteaching, or for the motivational scaffolding that struggling students often need. That work requires a person.

What makes Cosmo's model different from both platforms and conventional tutoring is where it starts: with diagnosis. Before instruction begins, Cosmo assesses exactly where a student's understanding breaks down — not just which topics they've covered, but which specific skills are solid and which are shaky underneath. That assessment shapes every session that follows.

From there, every class is live and 1:1 with the same teacher. Not a rotating roster of tutors working from a script. The same person every week, who builds a genuine understanding of how your child thinks, where they get stuck, and what kind of explanation actually lands for them. The research on consistent tutor relationships is clear: it matters.

Between sessions, Cosmo's app delivers practice that's directly tied to what was covered in class — feeding back into an AI-powered learning plan that adjusts as the student progresses. It's the combination that the research points toward: live instruction with a consistent teacher, supported by structured practice, all grounded in a real diagnosis of where the student is.

If you're not sure whether your child needs one-on-one support or whether an app might be enough, a single diagnostic session can tell you more than months of wondering. Cosmo's first trial class is free. Try a free class today →