Reach120 vs ELSA Speak: the 2026-format comparison, with its evidence attached
People ask whether it will help their Speaking score, and the honest answer needs the framing spelled out rather than a table. It reports a TOEFL-shaped number without being TOEFL preparation, which is not a criticism of it — it is what its own site says it is — and the two products solve different halves of the same complaint. This page separates them on the axes you can actually check, and says where its own evidence runs out.
Who wrote this, and how to read it
Read on 2026-08-24, without an account, buying nothing and accessing no paywalled surface. Free allowances and prices change faster than pages like this one do, so treat every figure as dated evidence rather than a current quote.
- There is no score, no rating and no verdict. There are no star ratings in this page or in its structured data — Reach120 has no published review corpus, so a rating would be a number we made up.
- What ELSA Speak does better is stated plainly, because a comparison that finds nothing good in the other product is an advertisement.
- Reach120 prices are not typed on this page. The pricing page is the only place they are stated, so there is one thing to keep correct rather than fourteen.
Reach120 and ELSA Speak, side by side
The spine of this table is the six task types the January 2026 format introduced, because coverage is the one axis with checkable evidence: either a public page names the task or it does not. Everything else — free access, metering, price, scoring transparency — is quoted.
| 2026 task type | Reach120 | ELSA Speak |
|---|---|---|
| Build a SentenceWriting | Runs it here | Not stated on the page we read |
| Write an EmailWriting | Runs it here | Not stated on the page we read |
| Academic DiscussionWriting | Runs it here | Not stated on the page we read |
| Listen & RepeatSpeaking | Runs it here | Not stated on the page we read |
| Take an InterviewSpeaking | Runs it here | Not stated on the page we read |
| Complete the WordsReading | Runs it here | Not stated on the page we read |
How the ELSA Speak column was filled: No cell is filled, and one of them deserves an explicit note because it is the cell most likely to be filled wrongly. ELSA’s own English homepage never names this exam at all: the exams it names are CEFR, IELTS and TOEIC. The TOEFL name appears only in its app-store description, and only as a mapping — an overall speaking score converted onto exam scales — never as practice of a task. Its interview coaching is general spoken-English interview practice; marking Take an Interview as named on the strength of the word "interview" would be us inventing a coverage claim the provider has never made.
| What you are comparing | Reach120 | ELSA Speak |
|---|---|---|
| Free access | Reading, Listening and Build a Sentence practice is free with no daily limit and no lifetime limit. Write an Email and Write for an Academic Discussion give a free account 3 AI-scored writing responses to start, then one more each day, up to 10 in total; Take an Interview has 3 AI-scored Take an Interview responses in total, for the life of the account; Listen & Repeat has 8 on the account, for life. Published questions, answers, transcripts, sample responses and PDF packs need no account. | Not stated in any form we can quote. The pricing page answered 404, and what is free against what is paid is described in an app-store listing rather than on the site — where it varies by storefront. |
| How AI use is metered | The free tier is metered by measured cost — the three tasks that spend a model on your own words carry daily allowances, Writing and Speaking counted separately because an Interview costs multiples of an essay. A paid pass carries a stated fair-use ceiling rather than an open-ended promise. | Not stated on the pages we read. |
| Price, as their page stated it | Exam passes are one-time and do not auto-renew; the current prices are on the pricing page, which this page reads rather than restates. | https://elsaspeak.com/en/pricing answered HTTP 404 on 2026-08-24, so ELSA publishes no rate card we could read. The only prices visible to us are in-app purchase tiers in an app-store listing, and those differ by storefront, currency and promotion — a figure that changes with the reader’s country is not one this page can state as a fact, so it states none. |
| Scoring transparency | How a score is produced, and what the language model is and is not allowed to do with it, is published at /methodology. No scorer-versus-human agreement figure is published: the internal benchmark corpus is too small to publish as an accuracy claim. | Its app-store description states an overall speaking score "mapped to major exams (IELTS, TOEFL, CEFR, TOEIC)". A mapping is a conversion, not a rubric score: nothing we read states how the underlying measurement is produced, and no comparison against real exam results is published. Reach120 publishes no agreement figure either — its benchmark corpus is too small to publish one honestly — and does publish how a score is produced, at /methodology. |
Every ELSA Speak cell above is a sentence their own public page stated, read on 2026-08-24. Prices and free allowances move; check the source before you buy. Reach120's own prices are not typed here — the pricing page is the only place they are stated.
What ELSA Speak does better
One thing, and nothing else in this comparison set states a pronunciation feature of any kind, ours included. It diagnoses pronunciation at the level of individual sounds and tells you which ones to move, which is the speech-motor problem that quietly caps a Speaking delivery score however good the content is. Reach120 can tell you delivery was the weak dimension in a response; it cannot tell you which vowel you are producing wrong, and no comparison table should hide that this product can.
Where Reach120 is different
They are not the same category, and treating them as one would misstate what ELSA says about itself. Reach120 practices the six task types the January 2026 format introduced, scores responses against the rubric and publishes its method at /methodology. ELSA trains pronunciation and reports an overall speaking score mapped onto exam scales, this one among them. Someone whose delivery is the problem and someone who has never answered a Take an Interview prompt need different things, and often need both.
Different is not better in the abstract. The difference matters only if it matches what you are short of — time, money, feedback, or lessons.
Where Reach120 is weaker, in its own words: This page makes no case for its own product over the others. Where Reach120 is weaker: there are no taught video lessons, no human tutor review of a response, no coverage of any exam other than the TOEFL test, fewer free AI-scored submissions a day than PrepEx states, and no published scorer-versus-human agreement figure — the internal benchmark corpus is too small to publish as an accuracy claim, so it is not published.
Which one to choose
- Choose ELSA Speak if what it does better is what you are short of. That sentence is above, unedited, and it is the whole reason this section can be brief.
- Choose Reach120 if you want to read the practice material and its answers before creating anything, and you want the four sections in one place with the method published.
- Choose both if the budget allows. They are not mutually exclusive, and nobody preparing for an exam owes one vendor loyalty.
Before either: Be clear about what you are buying, because a mapped number is easy to misread. ELSA reports a speaking score converted onto exam scales; that is a conversion of its own measurement, not a rubric score for a TOEFL Speaking task, and none of the 2026 task types is practiced in it. Its interview feature is general interview practice rather than the exam’s Take an Interview task. So if listeners cannot follow your pronunciation, this addresses that directly and nothing else in this set does; if you do not yet know how a 2026 task is answered, it is the wrong purchase. Prices are set per app store, so read the tier in your own storefront rather than any figure quoted elsewhere.
Sources, and what we could not read
- ELSA’s English homepage — https://elsaspeak.com/en/, read 2026-08-24.
ELSA Speak states the following about itself. These are their words, quoted, and this page does not verify them:
- Its app-store description states an "overall speaking score mapped to major exams (IELTS, TOEFL, CEFR, TOEIC)". That sentence is written by the vendor, and it is the only place this exam is named in anything of theirs we read.
Reach120 vs ELSA Speak: common questions
- Is ELSA Speak better than Reach120?
- This page does not answer that, and no page published by one of the two products honestly could. What it gives you instead is the evidence to answer it yourself: what each states about free access, price and 2026 task coverage, quoted from the page it was read from on 2026-08-24, plus a plain statement of what ELSA Speak does better — One thing, and nothing else in this comparison set states a pronunciation feature of any kind, ours included. It diagnoses pronunciation at the level of individual sounds and tells you which ones to move, which is the speech-motor problem that quietly caps a Speaking delivery score however good the content is. Reach120 can tell you delivery was the weak dimension in a response; it cannot tell you which vowel you are producing wrong, and no comparison table should hide that this product can.
- What does ELSA Speak cost?
- This page states no ELSA Speak price. https://elsaspeak.com/en/pricing answered HTTP 404 on 2026-08-24, so ELSA publishes no rate card we could read. The only prices visible to us are in-app purchase tiers in an app-store listing, and those differ by storefront, currency and promotion — a figure that changes with the reader’s country is not one this page can state as a fact, so it states none.
- Can I use Reach120 free instead of paying for ELSA Speak?
- You can use a large part of it free and decide afterwards. Reading, Listening and Build a Sentence practice is free with no daily limit and no lifetime limit. Write an Email and Write for an Academic Discussion give a free account 3 AI-scored writing responses to start, then one more each day, up to 10 in total; Take an Interview has 3 AI-scored Take an Interview responses in total, for the life of the account; Listen & Repeat has 8 on the account, for life. Published questions, answers, transcripts, sample responses and PDF packs need no account.
- Does ELSA Speak cover the 2026 TOEFL task types?
- Here is what we could establish rather than what we assume: No cell is filled, and one of them deserves an explicit note because it is the cell most likely to be filled wrongly. ELSA’s own English homepage never names this exam at all: the exams it names are CEFR, IELTS and TOEIC. The TOEFL name appears only in its app-store description, and only as a mapping — an overall speaking score converted onto exam scales — never as practice of a task. Its interview coaching is general spoken-English interview practice; marking Take an Interview as named on the strength of the word "interview" would be us inventing a coverage claim the provider has never made. A blank cell in the table on this page means we found no public statement, not that the feature is missing.
- Who publishes this comparison?
- Reach120 does, and Reach120 is one of the two products in it. That is disclosed at the top rather than buried, every competitor fact carries the URL and date it came from, and the section stating what ELSA Speak does better is mandatory on every page in this set. Nothing here is sponsored and ELSA Speak did not review it. If you believe a fact on this page is wrong or out of date, tell us and we will re-read the source.