Skip to content
Browsa
  1. Blog
  2. Captchas

Captchas, measured: only what each site's server accepted

Our captcha number is a count, not a rate: 216 of 223 attempts, each one accepted or refused by the site's own server, on production browsers, with every miss in the total.

Browsa team4 min readCaptchas

Most captcha claims are one percentage with no method behind it. Ours is a count you can check: 216 of 223 attempts across 13 captcha types, each accepted or refused by the server of the site that showed the captcha. This post explains how those attempts are run and counted, and where the numbers stop. The live table is on /captcha.

The numbers

These are the rows switched on for customers, as published on 17 November 2026. Each row is its own set of attempts on production browsers.

CaptchaAcceptedMedianExit
reCAPTCHA v2, 3 x 3 image grid16 of 1613.6 sData center
reCAPTCHA v2, 4 x 4 single image17 of 1816.1 sData center
reCAPTCHA v2, grid with reloading tiles14 of 1570.6 sData center and US ISP proxy
Image text, type the characters28 of 306.2 sData center
GeeTest v3, slide puzzle15 of 1612.9 sData center and US ISP proxy
GeeTest v4, slide puzzle15 of 1612.7 sData center and US ISP proxy
AWS WAF, captcha grid16 of 1615.1 sUS ISP proxy
hCaptcha, picture and drag puzzles16 of 1617.5 sUS ISP proxy
GeeTest v4, IconCrush (swap to make three)16 of 167.9 sData center
GeeTest v4, gobang (move one to make five)16 of 168.1 sData center
GeeTest v4, click the icons in order15 of 1614.1 sData center
GeeTest v4, pick the shape that matches16 of 1618.9 sData center
GeeTest v4, pick the photos that match16 of 1610.5 sData center
Total216 of 223

Three more checks pass with no solver at all, because the browser itself is accepted: reCAPTCHA v3 (score 0.9 on 5 of 5), Cloudflare's full-page check (5 of 5, one click) and Cloudflare Turnstile (15 of 15 tokens accepted by Cloudflare's siteverify on a widget we run for this purpose). They are not in the 216.

The reloading-tiles row is slow because of how that puzzle works: Google takes about 5 seconds to replace each clicked tile and asks for several rounds. Its one miss ran out of rounds at 134 s.

How we count

  • Production browsers. Attempts run on the same Phantom browsers and the same solver that customers use, not a lab copy.
  • A new challenge every load. Before a number is published, we check that the page serves a different challenge on each load and uses a real site key. A test key always passes, and a fixed picture can be memorized.
  • The site's server decides. A pass means the server behind the captcha accepted the answer: Google's verification on its demo, Cloudflare's siteverify, BotDetect's own check, GeeTest's server through its demo's login check, or AWS setting its WAF token and the page behind it loading. A token that the solver produced but no server accepted is not a pass.
  • Misses stay in. Each attempt uses a fresh browser. Every attempt that reached the captcha type counts, including the slow ones. When a solver is changed, the row shows the run of the version that ships, and earlier runs stay in its log.
  • Re-run monthly. The same benchmark runs again each month and the dated rows change with it. The script, capbench.js, and each row's attempt log can be downloaded from /captcha.

A type is switched on for customers only after at least 15 production attempts with at least 90% accepted. Press and hold (HUMAN) is at 11 of 12 and is marked Beta until it has more attempts.

We got this wrong once. An early Turnstile row of 5 of 5 came from a demo that runs Cloudflare's test key, which always passes. We replaced it with the server-verified 15 of 15 and wrote the correction on /captcha. That mistake is why the "real site key" check is in the list above.

Limits

  • Exit address. AWS WAF and hCaptcha were measured through a US ISP proxy only. From our data-center exit, the AWS test page answered 403 before any puzzle on 8 of 8 visits, and hcaptcha.com blocks the visit too. No solver can fix a visit that never reaches the puzzle.
  • Sample size. 15 to 30 attempts per type is small. 16 of 16 does not mean 100%: by the rule of three, the true miss rate is likely below about 19% (3 divided by 16), not zero.
  • Demo pages. Most rows use the vendor's own demo or a public demo page. A site can set its captcha harder than the demo.
  • Not covered yet. reCAPTCHA Enterprise passed 7 of 22 in tests and is not offered. DataDome was 0 of 5 on a protected site. Arkose is not measured: there is no public demo with a server verdict.

How claims usually look

Search for captcha solving and you will find "99% success" or "solves in seconds" on many landing pages. Few of them say which captcha, on which site, how many attempts, who decided what counted as a pass, from which IP address, or when. Without those, a percentage cannot be checked or compared.

If you are comparing providers, including us, ask for:

  1. The exact captcha type and variant, such as reCAPTCHA v2 4 x 4, not just "reCAPTCHA".
  2. Whether the site key was real and the challenge changed on each load.
  3. Whose server accepted each answer.
  4. Attempts and misses, not only a rate.
  5. The exit address type and the date of the run.

That is the format of every row on /captcha. When a number goes down at the next monthly run, the page will show the lower number.

Sources

  1. Browsa captcha results: every row, its demo recording and its attempt log
  2. capbench.js: the benchmark script
  3. Turnstile server-side validation, Cloudflare docs: how siteverify accepts a token
  4. Turnstile testing keys, Cloudflare docs: test keys that always pass
  5. Rule of three, a 95% bound when no misses are seen

Try it

Start free with 100 credits. No card.