In April 2026, Chegg told the SEC that its quarterly revenue had fallen 48% year over year, down to $63.3 million, with its core Academic Services business off 57%. Three years earlier the company was worth roughly $14.5 billion. Students had stopped paying for homework answers because ChatGPT and Google’s AI results give them away for free. Anyone weighing an online tutoring side hustle in 2026 probably read that headline and assumed the whole category was finished.
Four weeks later, a much smaller company filed a very different set of numbers. Nerdy, the parent of Varsity Tutors, reported first-quarter 2026 revenue of $48.7 million, gross margin up to 66.2% from 58.0%, and its second straight quarter of positive adjusted EBITDA. The detail worth sitting with: Nerdy’s active members fell 9%, but average revenue per member climbed 12% to $374 a month.
Read those two filings side by side and the story writes itself. AI destroyed the market for answers. It did not touch the market for a human being who shows up on a Tuesday and makes a 15-year-old actually do the problem set. That second market got smaller at the bottom and more expensive at the top, which is exactly the condition that makes a side hustle worth starting.
An online tutoring side hustle is paid one-to-one or small-group instruction delivered over video, usually 5 to 15 hours a week alongside a full-time job, at rates that ran roughly $19 to $40 an hour for general subjects in 2026 and $60 to $200 an hour for test prep and specialized technical subjects.
Last updated: August 2026
Quick answers
How much do online tutors make?
Online tutors typically earn $19 to $40 an hour for general academic subjects and $60 to $200 an hour for SAT prep, advanced math, and coding. Part-time tutors working 5 to 15 hours a week earn roughly $500 to $2,000 a month before platform fees. Full-time tutors at 30 to 40 hours reach $3,000 to $8,000.
Is online tutoring still worth it now that AI can answer homework?
Yes, for specialized instruction. AI wiped out the paid homework-answer business, which is why Chegg’s revenue fell 48% in Q1 2026. Live tutoring priced above that line grew more expensive in the same period, with Nerdy reporting revenue per member up 12% to $374 a month.
What do tutoring platforms take?
Wyzant retains a 25% platform fee on a tutor’s posted rate and bills students a separate 9% service fee. Preply charges new tutors 33%, sliding to 18% after 400 completed hours, and keeps 100% of the fee on every trial lesson with a new student.
How much do online tutors make in 2026?
Online tutors made $19 to $40 an hour for general subjects in 2026 and $60 to $200 an hour for test prep and technical subjects, with monthly income landing between $500 and $2,000 for someone working 5 to 15 hours a week. The number swings that widely because “tutor” describes two different jobs sharing one word.
The aggregators disagree in a way that’s useful once you know why. The Bureau of Labor Statistics, in its May 2025 occupational wage survey, counted 118,590 people employed under the tutor code at a median of $24.47 an hour and a mean of $28.39. ZipRecruiter’s online-tutor average runs closer to $18.80. Glassdoor’s sits near $40. Indeed lands at $27.67.
None of them are wrong. ZipRecruiter is scraping job listings, which skews toward the high-volume English-conversation and homework-help roles that platforms post constantly because turnover is brutal. Glassdoor is weighted toward self-reported earnings from people who bothered to report, which skews toward the ones doing well. The BLS number is the honest middle for W-2 tutoring work, and it excludes almost everyone running this as an independent side hustle.
Here’s what matters for your own math. Rate is a function of two things: whether the student can get the same outcome from ChatGPT, and whether anyone is checking your work. Conversational English tutoring fails both tests and pays $15 to $20 an hour. SAT prep passes both, because a score report is a verifiable outcome and a nervous parent is the buyer, and it pays $45 to $200.
| Hours per week | At $20/hr | At $45/hr | At $90/hr | Realistic for |
|---|---|---|---|---|
| 5 hours | $433/mo | $975/mo | $1,950/mo | Two weeknights, no weekend |
| 10 hours | $867/mo | $1,950/mo | $3,900/mo | Weeknights plus Saturday morning |
| 15 hours | $1,300/mo | $2,925/mo | $5,850/mo | Near the ceiling with a day job |
Those are gross figures at 4.33 weeks a month. Marketplace fees take 18% to 33% off the top, and roughly a third of your unpaid time goes to scheduling, prep, and parent emails. Budget accordingly. Tutoring consistently outearns the campus-job tier in GJ’s ranking of side hustles for college students, and the reason is the same one driving the spread above: subject expertise is the only input nobody can commoditize.
Why tutoring pay rose while Chegg fell apart
The AI disruption in education split one market into two, and only one half died. Chegg sold answers at scale. Varsity Tutors sells a person. Generative AI is very good at the first thing and structurally bad at the second, and the 2026 filings make the split legible in a way that opinion pieces don’t.
Chegg’s Q1 2026 10-Q put total net revenue at $63.3 million, down 48%, with Academic Services down 57%. The mechanism was mundane. A student used to type a calculus problem into Google, click through to a Chegg answer page, and hit a paywall. Google’s AI now prints the answer above the results. The click never happens, the funnel starves, and 99% of a $14.5 billion market cap goes with it.
Nerdy’s Q1 2026 8-K shows the other half. Revenue of $48.7 million, up 2%. Gross margin at 66.2%, up from 58.0%. Adjusted EBITDA positive at $1.0 million against a $6.4 million loss a year earlier. Active members down 9% to 36,900, and average revenue per member per month up 12% to $374. The company reaffirmed full-year guidance of $180 million to $190 million.
Fewer customers, each paying meaningfully more. That’s not a company in decline, it’s a company whose cheap segment left for free tools while its expensive segment stayed and absorbed price increases. Nerdy said as much: the margin expansion came from prior price increases and AI-enabled efficiency in its own operations.
For anyone deciding whether to start, the practical read is this. Every hour of tutoring you can imagine being replaced by a well-prompted chatbot is an hour that will keep getting cheaper. Every hour that depends on accountability, diagnosis, a credential, or a parent’s willingness to pay for reassurance is an hour with pricing power. Pick the second kind. The same sorting logic is showing up across this year’s side hustle trends, where skill-gated work held its rates and commodity gig work didn’t.
Which subjects pay the most
Test prep, advanced STEM, and professional certification prep pay the most, generally $60 to $200 an hour, while conversational language practice and general homework help sit at the bottom near $15 to $25. The gap tracks how measurable the outcome is and how motivated the payer is.
SAT and ACT prep is the best-documented tier. Entry-level and graduate-student tutors charge $45 to $80 an hour. Experienced independent tutors charge $80 to $200. Section specialists inside structured programs charge $150 to $500 and up, and in New York, San Francisco, Los Angeles, Boston, and Washington DC the private-market floor sits around $100. Families routinely spend $1,500 to $6,000 on a 20 to 30 hour program, which tells you the budget exists before you ever quote a number.
| Subject | Typical rate | What you need | AI exposure | Best for |
|---|---|---|---|---|
| SAT / ACT prep | $45–$200 | Your own score report | Low | Anyone who tested well in the last decade |
| Coding and data science | $60–$150 | Shipped work, a GitHub profile | Medium | Working engineers with two free evenings |
| AP and college STEM | $40–$90 | Coursework in the subject | Medium | Grad students and recent grads |
| Professional certifications | $50–$120 | The certification itself | Low | Accountants, nurses, project managers |
| K-12 reading and math | $25–$55 | Patience, a background check | Medium | Teachers moonlighting |
| Conversational English | $15–$25 | Fluency and a webcam | High | Volume, not rate |
One counterintuitive note on that bottom row. Conversational English is the easiest category to enter and the one most exposed to AI voice tools, so treat it as a way to log hours and reviews rather than as the business itself. Preply’s commission tiers reward accumulated hours, which means cheap early lessons genuinely do buy you a better take rate later.
What the platforms actually take from your rate
Marketplace platforms take 18% to 33% of a tutor’s rate in 2026, and the difference between a good and bad platform decision is worth thousands of dollars a year. Most guides skip this section. It’s the one that determines whether the side hustle is worth your Tuesday.
Wyzant, now a division of IXL Learning, is the simpler of the two majors. Its tutor fee documentation states that tutors retain 75% of their posted hourly rate and Wyzant keeps a 25% platform fee. Students pay an additional 9% service fee on top. There’s a real loophole worth knowing: when a tutor refers a brand-new student who then books with them, that tutor receives 100% of their rate with no platform fee deducted. If you already have a few people who’d hire you, route them through your own referral link and you keep the full amount while Wyzant handles payments and scheduling.
Preply runs a tiered model documented in its commission help article. New tutors pay 33%. That falls to 28% after 20 taught hours, 25% after 50, 22% after 200, and 18% once you cross 400. Preply also keeps 100% of the fee on every trial lesson with a new student. On a $10 rate, a brand-new tutor nets $6.70 and a 400-hour veteran nets $8.20 for the same lesson.
Run the numbers on a realistic year. Ten hours a week at $50 is $26,000 gross. On Preply at the new-tutor 33% rate you keep about $17,400. At the 18% veteran rate you keep $21,300. Off-platform, invoicing directly, you keep the whole $26,000 minus roughly 3% in payment processing, so about $25,200. The platform is buying you clients, and in year one that’s a fair trade. By year two it’s an expensive customer acquisition cost for people who already know your name.
The move most tutors get wrong is treating the platform as the destination. Use Preply or Wyzant to source the first 8 to 10 students and build a review history, then run renewals and referrals directly. Check the terms of service before you do anything that looks like circumvention, because both platforms prohibit poaching an existing match. Referrals from a happy parent to their neighbor are a different thing entirely, and that’s where the durable book of business comes from. The same graduation path shows up in turning a skill into an online business.
How do you start a tutoring side hustle in 2026?
Starting takes about a week and under $100. The steps are narrow the subject, price against the local market, list on one marketplace, run three sessions free to build reviews, then raise the rate. Most people fail at step one by offering everything to everyone.
1. Pick one subject and one buyer. “Math tutor” competes with 40,000 listings. “Algebra 2 for students who failed the first semester” competes with almost none, and the parent searching that phrase is ready to pay today. Specificity is pricing power. If you scored well on a standardized test, that score report is your entire credential.
2. Price at the local market, not the platform average. Look up what an in-person tutor charges in your metro and quote 70% to 80% of it. Tutors who anchor to the ZipRecruiter national average of $18.80 leave real money behind, because the parent hiring you is comparing against a $90-an-hour local center, not against a marketplace median.
3. Set up the technical minimum. A wired connection, a USB microphone in the $60 range, and a shared whiteboard. Preply and Wyzant both provide their own classrooms. If you’re going direct, Zoom plus a tablet and stylus is enough. Nobody has ever hired a tutor for their lighting setup, but bad audio ends relationships fast.
4. Trade three free or discounted sessions for reviews. Marketplace ranking is driven by review volume and response time. Three strong reviews in week one moves you off page four. Preply’s structure already assumes this since it takes the full fee on trial lessons.
5. Raise your rate every 20 students, not every year. Rate increases are the only lever that scales without more hours. Give existing clients 30 days’ notice, keep them at the old rate for a term, and quote new inquiries at the new number. Nerdy raised prices into a shrinking member base and expanded gross margin by 820 basis points. The same logic works at a scale of one.
6. Handle taxes from the first dollar. This is 1099 income. Set aside 25% to 30%, track mileage if you do any in-person work, and deduct the microphone, the subscription tools, and a share of your internet bill. Doing this at the start avoids a bad April. If you’re keeping a full-time job alongside it, GJ’s guides to growing a side hustle without quitting and building one around a day job cover the scheduling side.
Is online tutoring a good side hustle?
Online tutoring is a good side hustle for people with a verifiable subject credential and 5 to 15 predictable weekly hours. It’s a poor one for anyone hoping for passive income, because every dollar is tied to an hour you personally sit through.
Set it against the baseline. Bankrate’s 2025 survey found 27% of US adults had a side hustle, down from 36% the year before, with median monthly earnings of $200, down from $250. Most side hustles are not clearing meaningful money. Ten tutoring hours a week at $45 gross beats that median in a single weekend, which is the entire argument. Set against the gigs people actually pick for flexibility, tutoring trades some scheduling freedom for a much higher floor.
The honest downsides are real. Demand is seasonal, with test-prep peaks before spring and fall exam dates and a dead zone in July. Cancellations are constant and most platforms cap what you can recover. It doesn’t compound: stop teaching and income stops that week, unlike the creator-platform models where a back catalog keeps earning. And you’re doing cognitively demanding work in the evening, after a day of cognitively demanding work.
Three free hours. That’s what a 5-hour week really costs once you count scheduling and prep. Decide whether the money is worth them before you list a profile.
Turning tutoring into an actual business
The transition from tutor to tutoring business happens at one of three places: group instruction, a productized program, or subcontracted tutors. Each one breaks the hour-for-dollar ceiling in a different way, and only the first is available immediately.
Small groups. Four students at $30 each is $120 an hour for the same prep you already did for one at $50. Parents accept it because the per-seat price drops and their kid gets peers. This is the single biggest rate increase most tutors never take, and it requires no new skill beyond running a room.
A named program. Package eight weeks into “Algebra 2 Recovery” with a diagnostic, a defined syllabus, and a progress report, then sell it for $1,200 instead of billing $50 an hour sixteen times. Buyers pay for outcomes and structure, not for time. This is precisely what lets structured programs charge $150 to $500 an hour for a section specialist while an unbranded independent charges $80.
Subcontracting. Once inbound demand exceeds your calendar, hand overflow to another tutor at 60% to 70% of the rate. Margins are thin and quality control becomes your job, so most solo tutors should stop at the first two. Varsity Tutors and Wyzant are both this model at scale, and Nerdy’s 66.2% gross margin is roughly what the economics support with real infrastructure behind them.
There’s a fourth path worth naming, because 2026 made it viable. Record the program once, sell the recording alongside a monthly live Q&A, and the marginal cost of the tenth buyer approaches zero. That’s the same structural shift that’s reshaping AI-native content businesses, applied to a subject you already teach.
The takeaway from Chegg and Nerdy holds all the way down to a one-person operation. Selling information is a losing position in 2026. Selling accountability, diagnosis, and a person who notices when a student stops showing up is a business that AI made scarcer and more valuable at the same time. Online tutoring is the rare side hustle where the technology shift argued in favor of the human.
