AMP vs Brilliant

Brilliant builds its lessons from interactive problems, with apps, an AI tutor and coding courses1,10, for less than AMP costs2. Its math starts near grade 4, and we found no stated mastery rule or review schedule in its published materials7,9,11. AMP runs from counting to precalculus, chooses every next problem, and brings each learned skill back on a fixed schedule.

Try it free

Why AMP

AMP starts at counting, not at fractions

Brilliant's math begins with fractions at about grade 4 or 5 and runs to the opening topics of calculus7. Its help center says it "does not present a complete K–3 arithmetic sequence"7. Its free K–5 practice library held kindergarten and grade 1 quizzes when we checked, with grades 2 to 5 marked coming soon20. AMP covers counting through precalculus on one map of 336 skills, so a learner who struggles with fractions is sent back to the skills fractions build on, however early they sit.

The research: Practice at the learner’s level

A written rule for when a skill is learned

Each Brilliant concept ends with a skill check, and unit checks reach back to earlier material11. We found no stated threshold for mastery in its published materials11. In AMP a skill counts as learned when four of the last five answers are right without help, and an answer given after a hint or the lesson does not count. Two misses in a row send practice back to what the skill builds on, and the rule sits in public code anyone can read.

The research: Mastery, and repair of gaps · The research: One-step hints

Old skills come back on a fixed schedule

Brilliant says important ideas return across sessions, and it names spacing and interleaving among its methods9. Missed problems on a unit check add review lessons to the plan11. We found no schedule of review intervals in its published materials9. AMP brings each learned skill back after 6 hours, 1 day, 3 days, 1 week, 2 weeks and a month, mixed with other topics, and a miss starts the schedule over.

The research: Spaced, mixed review

A worked example behind every problem

Brilliant's lessons often pose a question before showing the procedure, then connect visual models to the notation9. In AMP every one of the 1,384 problem types has a written lesson with a worked example, and worked steps follow every answer, right or wrong. Across 55 studies, worked examples improved math performance by an average of g = 0.4824.

The research: Lessons with worked examples · The research: Feedback after every answer

Paper workbooks, and records a homeschool can keep

Brilliant needs an internet connection, and its lessons are not available offline6. It tells homeschool families to build their own portfolio and records8. Its parent dashboard comes only with Family or Single Child plans, shows activity and accuracy on checks, and says it is not a grade13. AMP prints unlimited workbooks made for each learner, with worked answer keys, and gives the adult a level in every strand, progress reports and report cards.

The research: Progress parents and teachers can see

No AI model writes what a learner sees

Koji, Brilliant's AI tutor, is built partly on a third-party language model and is designed to ask guiding questions rather than hand over answers10. AMP runs no AI model while a learner works. Each hint is one step built from the problem's own numbers, a test checks that no hint gives the answer, and 185 independent checkers re-derive the answers.

The research: One-step hints

Side by side

AMPBrilliant
Price$599 a year or $59 a month for the first child, then $399 or $39 for each more. Two parents practice free. Full refund within 30 days.Monthly: Individual $40, Family (six seats) $60 a month. Annual, billed yearly: Individual $14, Family $28 a month, Black Friday prices 30% off $20 and $402. Free tier3.
Who can buyFamilies, tutors and schools, directly. Free to start without a card: 100 answers and 5 workbooks. Arizona ESA, Arkansas EFA and New Hampshire EFA funds pay for it.Families and individuals, online or through app stores2. Trials need a card4. ESA funds in Arizona, Arkansas and Florida5.
RangeCounting through precalculus in one program: 336 skills on one map.From fractions at about grade 4 to opening calculus topics, plus coding and science1,7. No complete K–3 arithmetic sequence7.
Finding the starting pointNo separate test. Placement takes the first 24 problems of practice. After that, every answer adjusts a level in each of 14 strands of math.A short diagnostic at signup picks a starting topic. Jumping ahead takes a passed level check12.
TeachingA one-step hint and a written lesson with a worked example for every problem type. Worked steps after every answer. Narrated lesson films.Interactive problems: move a point, test a pattern, pick a strategy9. Koji, an AI tutor, asks guiding questions10.
What decides the next problemAMP chooses every problem from the learner’s answers: new work just within reach, a prerequisite, or review.A recommended topic order; Premium learners may take topics in any order19. Practice uses learning history to pick problems9.
MasteryFour of the last five answers right without help. An answer after a hint doesn’t count. Two misses in a row send practice back to what the skill builds on.A skill check ends each concept. No stated mastery threshold found in its published materials11.
ReviewScheduled. Learned skills come back after six hours, a day, three days, a week, two weeks and a month, mixed with other topics.Ideas return across sessions; missed unit-check problems add review lessons9,11. No schedule of intervals found9.
PrintUnlimited workbooks made for each learner, each with a worked answer key. Results entered from paper count.Lessons need an internet connection and are not available offline6. The free K–5 library mentions printable worksheets in a guide20.
What the adult has to doNothing to plan, teach or grade. AMP places the learner, chooses each problem, checks it and schedules review.None required. Built for learners who work independently18. Homeschool families keep their own records8.
ReportsThe level in every strand with its likely range, skills, time, and the actual problems behind each number. For teachers, who needs them today.Parent dashboard on Family or Single Child plans: days active, problems solved, accuracy on checks. Not a grade, it says13.
EvidenceA published rationale (ESSA Tier 4) with the research behind each part, public source code, and outcome studies from 2027, each registered before it starts.No outcome study found on its site, in Evidence for ESSA or WWC intervention reports21,22. Its Koji figures are behavioral signals, not a controlled study18.

What Brilliant is

Brilliant teaches math, coding, data science and science through short interactive lessons, on the web and in apps for iPhone and Android1,16. Its math runs from fractions at about grade 4 to the opening topics of calculus7. It says it is trusted by more than 5,000 schools1.

A learner opens a topic and works through short concepts built from interactive problems: move a point, test a pattern, estimate, choose a strategy9. Each concept ends with a skill check. Unit checks cover the unit and earlier material, and use a map of prerequisite skills to find the gap behind a miss11. A short diagnostic at signup picks the starting topic12. Koji, an AI tutor drawn as a character, can see the problem, ask guiding questions and sketch on the interactive10. Streaks and XP track the habit17.

Brilliant describes itself as a fit for learners who work independently18. It says it is not a complete K–12 school, and that it can serve as the core for math with other support in the early grades8. A free tier unlocks two concepts or practice sets a day3.

What Brilliant does well

The evidence

Brilliant

We found no study of Brilliant's effect on math achievement on its website, in Evidence for ESSA, or among the What Works Clearinghouse's intervention reports, as of November 202621,22.

Brilliant's own numbers come from a July 2026 analysis of middle and high school learners who asked Koji for help. 94% tried again, 66% then answered correctly without Koji giving the answer, and 87% of those who moved on got the next problem right18. Brilliant calls these behavioral signals and says they are not a substitute for a controlled academic study18. The analysis describes no comparison group, and the figures describe behavior inside the product, not test scores.

Its FAQ says interactive learning is "6x more effective" than passive methods, and links a news article; the page describes no study of Brilliant6.

AMP

Brilliant and AMP lean on some of the same science: Brilliant names spacing and interleaving9. AMP turns them into a rule. In a randomized trial with 787 seventh graders that meets What Works Clearinghouse standards without reservations, mixed and spaced practice scored 62% on the researchers' own test a month later, against 38% for practice in blocks23. Worked examples, which AMP attaches to every problem type, average g = 0.48 across 55 studies24. We expect a program that applies both at every problem to show it in outcomes, and we have committed to measuring that in public.

AMP is new, and its own outcome studies have not run yet. They begin in 2027: growth on a national test against hours of use, then comparison groups, then randomized trials. Each will be registered before it starts and published whatever it finds. Until then, the research behind each part of AMP is laid out in our report, and the code is public.

The research behind AMP

Choose Brilliant if

Choose AMP if

Questions

Is Brilliant a complete math curriculum?

Brilliant says it is not a complete K–12 school, and that early elementary learners may need other support8. It does not present a complete K–3 arithmetic sequence7. AMP runs from counting through precalculus in one program.

Is AMP more expensive than Brilliant?

Yes. AMP is $599 a year or $59 a month for the first child, and $399 or $39 for each more, with two parents practicing free. On November 21, 2026, Brilliant's pricing page showed Individual at $40 a month on the monthly plan, or $14 a month on the annual plan, billed as one payment. Family, with six seats, was $60 a month, or $28 a month on the annual plan. The annual prices were a Black Friday sale, 30% off $20 and $402.

Does AMP have an AI tutor like Koji?

No. AMP runs no AI model while a learner works. Every problem type has a one-step hint that never gives the answer, a written lesson with a worked example, and worked steps after each answer.

Can we use AMP and Brilliant together?

Yes. Some families may use AMP for the core path and its review schedule, and Brilliant for interactive puzzles or coding. AMP's first 100 answers and 5 workbooks are free, with no card.

Has Brilliant been shown to raise math scores?

We found no outcome study of Brilliant on its site, in Evidence for ESSA or in the What Works Clearinghouse's intervention reports21,22. Its published figures track behavior inside the app, and Brilliant says they are not a substitute for a controlled study18.

Sources

  1. Brilliant (home page)
  2. Brilliant Premium pricing (Brilliant)
  3. What's the difference between Free and Premium? (Brilliant Help)
  4. Does Brilliant offer any free trials? (Brilliant Help)
  5. Ways to pay for Brilliant (Brilliant Help)
  6. Frequently asked questions (Brilliant)
  7. What math does Brilliant cover? (Brilliant Help)
  8. Can I use Brilliant for homeschooling? (Brilliant Help)
  9. How Brilliant teaches math (Brilliant)
  10. How does Koji work? (Brilliant Help)
  11. Skill checks and unit checks (Brilliant)
  12. Course placement guide (Brilliant Help)
  13. Parent progress dashboards on Brilliant (Brilliant Help)
  14. Is Brilliant accredited? (Brilliant Help)
  15. Brilliant for educators (Brilliant)
  16. Where can I use Brilliant? (Brilliant Help)
  17. Does Brilliant offer certificates? (Brilliant Help)
  18. Is Brilliant worth it? (Brilliant)
  19. What are learning paths? (Brilliant Help)
  20. Free K–5 math practice (Brilliant)
  21. Evidence for ESSA, site search for Brilliant
  22. What Works Clearinghouse, intervention reports
  23. Rohrer, Dedrick, Hartwig & Cheung (2020), A randomized controlled trial of interleaved mathematics practice (What Works Clearinghouse review)
  24. Barbieri et al. (2023), A meta-analysis of the worked examples effect on mathematics performance

Checked November 2026.

Brilliant is a trademark of its owner. AMP is not affiliated with it. Found an error? Write to skylar@corpora.inc.

More comparisons

All comparisons · AMP · Research