How to choose an abacus program: the six questions that actually decide it
Centre, live online class, or self-paced app — the delivery model decides more than the brand does. Six questions, answered with what each provider actually publishes, including us.
Parents ask which abacus program is best. What actually gets decided is narrower: which delivery model your family can sustain for two or three years. The brand, the bead colour and the certificate all follow from that. One disclosure belongs at the top of a page like this — Kani sells a self-serve abacus app, so we are an interested party, and in this category every comparison page that ranks was written by somebody with something to sell. This one included.
The tool, the method and the brand are three separate choices
An abacus is hardware. The modern soroban carries one bead above the reckoning beam worth five and four beads below worth one each, so a single rod holds any digit from 0 to 9 — that is the whole instrument, and buying one is a single purchase rather than a subscription. The method is what a child does with it: bead technique, then the five-complements and ten-complements, then the same work with the frame taken away. The named programs are the third thing — franchise networks that license a curriculum built around that tool to local centre owners. So when you compare programs, you are mostly comparing businesses rather than arithmetic.
Question 1: how is it delivered?
| Model | Schedule as published | Who sets the price | Strength | Where it strains |
|---|---|---|---|---|
| Franchise centre | One long class a week is the common pattern: a SIP Abacus franchisee publishes one weekly session of 2 hours 15 minutes with a maximum of 15 students, and ALOHA’s Bangladesh master franchise publishes one two-hour class a week. | The local centre owner. Franchise fees are set locally, which is why almost no network publishes a national price. | A teacher in the room, a cohort, and an appointment that is hard to skip. | Long sittings for young children, and you cannot check the fee before you enquire. |
| Live online class | Mastermind Abacus publishes its online programme as twice a week for one hour each, at a 1:5 teacher-to-student ratio. | The provider — though that page publishes no fee either. | Small-group attention with no travel, and a published ratio, which is rare. | Still schedule-bound across time zones, and the daily practice happens off-camera. |
| Self-paced app | You choose the slot. SoroTouch states it runs on tablets only, not on smartphones or computers; Kani is built around short daily sessions. | The vendor, in public: Abacus4Everyone lists INR 399/month, Vedaavi lists USD 2.49/month, Kani about USD 7/month with a free tier. | The price is visible before you commit, and practice can be daily rather than weekly. | No cohort and nobody to notice when you stop; technique has to be checked by the software. |
Read those schedules against your own child rather than against each other. One abacus teacher, writing on her teaching blog in July 2024, describes classes that "usually run for about 75 minutes, which can be quite a stretch for young kids", and puts five- and six-year-olds at "about 10-15 minutes before they begin chatting, getting restless, or asking for a break". That is one practitioner describing her own students — no retention data behind it, and no bearing on any particular provider — but it is the right thing to watch for at a trial lesson.
The other number worth holding on to is the homework. A SIP Abacus FAQ post, published 21 January 2025, pairs weekly classes "lasting around two hours" with home practice of "approximately 10 minutes" a day. Every model in that table rests on those ten minutes, and in all three cases they happen at your kitchen table.
Question 2: what will the whole ladder cost, not the first month?
Ask for the cost of the full ladder, in your currency, before you enrol in level one. Most networks publish no figure and direct you to a local centre, because franchise fees are set locally: SIP Abacus’s FAQ says tuition "can vary by location and specific course offerings" and that it is best to contact your local centre; Kumon UK’s FAQ says "the registration and monthly fee varies slightly from centre to centre"; SUPERA in Brazil runs a page headed "Qual é o valor da mensalidade do Supera?" which answers with the same explanation and points readers to a nearby unit. That is a distribution model, not a secret. The published exceptions make useful anchors. ALOHA Ireland lists "ALOHA Mental Arithmetic costs 420€ / Term (Level)", payable in full or in four equal instalments of 105€, for two hours of class a week in groups of "a maximum of 10 students", with 17-week terms and two terms a year (as listed, retrieved July 2026) — and that fee buys live in-person teaching, so it is not like-for-like against software. Indian Abacus publishes coaching from "INR 1800/per Level for 3 Months" plus a one-time INR 1,100 for materials, and notes that rates vary by centre location tier (as listed, retrieved July 2026).
Question 3: how many levels, and how many years?
- Brainobrain’s own site: a main programme of 10 levels at 3 months each for ages 7-14, plus a LittleBobs track of six three-month modules for ages 4-6.
- A SIP Abacus franchisee’s course page: eleven levels grouped Foundation (4), Advance (4) and Grand Master (3), at 13-16 weeks per level.
- Mastermind Abacus online: 8 levels of 3 months each.
- ALOHA’s Bangladesh master franchise: Junior, six levels over 18 months for ages 4 to under 7; Senior, eight levels over 24 months for ages 7 to under 14, with Basic, Master and Grand Master certificates after levels 3, 6 and 8.
- Abacus4Everyone, a self-serve platform: 14 levels — the eight core belts from Yellow to Black plus advanced levels 9-14 (as published July 2026).
- Kani: nine levels, Foundation to Grand.
Do the multiplication before you fall for a brand. Ten levels at three months each is two and a half years of continuous enrolment; eight levels at three months is two years. Rung counts are also not comparable between providers — a ladder cut into eight steps is not shorter than one cut into fourteen, it is cut differently — so compare months of instruction and hours per week instead. And the research promises no quick return: Chunjie Wang’s 2020 review in Frontiers in Neuroscience notes that most studies detecting a benefit did so after around three years of training.
Questions 4 and 5: the right age, and what the evidence supports
Published starting ages vary so widely that age is plainly not the deciding variable. SoroTouch gives "5 to 8 years old" as its recommended start. Mastermind Abacus’s online page says "children aged 7 to 12 years" in body copy while its registration form says "6 to 12 Years old". A SIP Abacus FAQ describes the programme for children aged 6 to 12. SUPERA in Brazil says it is for everyone from age 5 with no upper limit. Kani treats level as an explicit placement rather than something derived from age, which is the more useful frame for a nine-year-old beginner or an adult one: the question is which technique is missing, not which birthday has passed.
Then check what a program claims against the strongest studies available. A randomized controlled trial in Vadodara, India (Barner and colleagues, Child Development, 2016) enrolled 204 children, 183 of whom completed, at three hours a week: mental-abacus students outperformed controls on arithmetic, Cohen’s d = .60 with a 95% confidence interval of .30 to .89 on the study’s own measure, but the training "did not alter basic cognitive abilities", and baseline spatial working memory moderated how much each child learned — children below the median did not differ from controls. A one-year classroom-randomized trial with 180 children (Barner and colleagues, Journal of Numerical Cognition, 2017) saw no differential change on either its in-house arithmetic measure or the standardized WJ-III, and concluded: "Overall, our results suggest caution in the adoption of MA as a short-term educational intervention." Wang’s 2020 review reports potential gains in mathematics, working memory and numerical magnitude processing, but says the effect on fluid intelligence "appears to not go beyond a placebo effect". A 2024 PLOS ONE trial of one tablet app, SoroTouch, is often quoted: it has 20 participants, ten per arm, and its funding statement discloses that the app’s developer funded the work and had a role in data collection. The practical rule that follows is easy to apply on any sales page, ours included — a program that promises higher intelligence scores or dramatic brain changes is describing its marketing, not this literature.
Question 6: what if you stop, switch, or have no centre nearby?
Plan both exits before you pay. Switching mid-ladder: level names do not port. Belts, grades counted downward from Grade-10 to Grade-1, and Foundation-Advance-Grand Master groupings are different cuts of the same technique, so a move is a re-placement rather than a transfer of credit. If you want a yardstick nobody is selling you, brand-independent ones exist: in Japan the proficiency exam is split between the Japan Chamber of Commerce and Industry and the Japan Abacus Federation, with grades 1-3 requiring 240 of 300 points to pass, while the League for Soroban Education of Japan runs a parallel kyu-and-dan ladder plus a formally graded flash mental-arithmetic exam; PAMA Global in Taiwan publishes sample papers for levels 1 to 12 plus Advanced. The second exit is having no centre within reach. Kani has no physical locations anywhere, so we will not pretend otherwise: what you give up without a centre is the cohort and the fixed appointment. What you do not give up is the technique or the competition circuit, which associations and ministries run independently of the schools — Iraq’s Ministry of Education announced on 23 February 2025 that Iraq had taken first place for the second consecutive year in the Arab Memory and Mental Arithmetic Championship, which it said drew 500 Arab participants from 12 countries, and Qatar’s Ministry of Education and Higher Education ran a first national mental-arithmetic competition for Grade 1 and Grade 2 students across government schools, reported in early May 2024.
Where Kani fits, and how to place a child who started elsewhere
Stated plainly so you can rule us out quickly: Kani is a self-serve app with nine levels from Foundation to Grand, about USD 7 a month with a free tier and checkout you complete yourself, eight site locales, native Android and iOS apps, voice-answered drilling for children who can say an answer faster than they can type it, and a competition simulator built on 200-question category papers that print with a matching answer key. Where that sits on price depends entirely on what you compare it against: it is far below live centre tuition such as ALOHA Ireland’s 420€ per term-level, and above self-serve apps such as Abacus4Everyone at INR 399 a month or Vedaavi at USD 2.49 a month. One thing worth saying out loud about everyone in this category, us included: no operator publishes audited figures for students taught or centres open, so treat every footprint number you read — ours too — as marketing. And if your child already finished a level or two elsewhere, drop the old label and test the technique in order: counting on the beads, five-complements, ten-complements, multi-row addition, then multiplication and division. The first step that wobbles is the rung to start on.
ALOHA, Brainobrain, SIP Abacus, Mastermind Abacus, Kumon, SUPERA, SoroTouch, Abacus4Everyone, Vedaavi and PAMA Global are trademarks of their respective owners. Kani is not affiliated with, endorsed by, or licensed by any of them. Every figure above is quoted as published on the date given and may change without notice.