PassGen.co

Language

Password Strength Checker

Your password never leaves your device. How it stays private

Try:

  • See how long it would take to crack, and whether it has leaked.

Password manager

How Secure Is Your Password?

Type or paste a password above to see how strong it is, how long it would take to crack, and how it would be cracked, and whether it is one of the million most-leaked passwords. It all happens in your browser as you type, and the password never leaves this device.

When you stop typing, it also looks the password up in the Pwned Passwords list from Have I Been Pwned, sending only 5 characters of a scrambled fingerprint, never the password, and you can watch it happen. See exactly what is sent.

Jump to: What the result means · What to do · How passwords are cracked · Time to crack · Is it safe? · The future · Questions

What the Result Means

Your result is the time 1,000 graphics cards would take, on average, to guess the password from a stolen copy of a website’s password list. That time sets the level:

Very strong
100 years or more
Strong enough for important accounts, as long as it is used only there and has not leaked
Strong
Less than 100 years
Good for most accounts, if unique
Fair
Less than 10 years
OK for less important accounts, not for email or banking: a bigger attacker or faster hardware would need far less
Weak
Less than a day
Change it now, everywhere you use it

Under Time to crack, the same password also appears where guessing is slow by design: Wi-Fi, and a password manager or encrypted drive. So a password can be Weak for websites and still fine for your Wi-Fi. For a password a person made up, these two times are 1,000 times shorter than the plain estimate, because cracking tools are better at people’s patterns than any estimate can be.Pick 1 GPU or 100,000 GPUs to see the times with less or more computing power; the result always uses 1,000.

“How it would be cracked” names the method that sets the time and links to how it works. A password found in a data breach is unsafe whatever its strength, because attackers try leaked passwords before anything else. A checker can show that a password is weak; it cannot prove that one is safe.

What to Do If It Is Weak or Leaked

  • Change it on every site where you use it, starting with your email, because email resets everything else.
  • Make a new one rather than tweaking the old one. Adding a character or changing 1 to 2 is the first thing attackers try. Make a random password or a passphrase.
  • Don’t build it from things people guess: names, birthdays, years, sports teams, keyboard runs like qwerty, look-alike swaps like @ for a, or a word with 123 or ! on the end.
  • Use a password manager, so every account can have its own password without you remembering any of them.
  • Turn on two-step verification wherever it is offered. It helps protect an account even if the password leaks.

How Passwords Are Cracked

Hardly anyone cracks a password by typing guesses into a sign-in page, because sites block repeated wrong tries. The real danger is a breach: attackers copy a site’s stored passwords and guess against that copy on their own computers, as fast as their hardware allows. They don’t guess at random. Cracking tools work through the likeliest passwords first, in roughly the order below.

Looks Strong, Isn’t

Many simple checkers count length and kinds of character, so they rate a password like P@ssw0rd2024! highly. Cracking tools don’t count. They try the patterns people use, and they try them first. Four passwords that look alike, and what this checker makes of them:

  • Password123!LeakedOn the lists of leaked passwords, tried first
  • OrangeSunsetBeachWeakThree common words, found by a dictionary attack
  • Sunfl0wer2024!WeakA word with a look-alike swap, a year, and a symbol: the usual changes
  • k7#Rq!9vT$2m@LxVery strongRandom, so only trying every combination works

1. Leaked passwords

First come passwords real people have used, collected from earlier breaches: hundreds of millions of them, the most popular first. A password on those lists falls in seconds however strong it looks, which is why this checker looks it up before anything else.

2. Words and names: a dictionary attack

Next come word lists: dictionaries in many languages, first names and surnames, places, sports teams, song lyrics, and words joined together. Common words go before rare ones, so a passphrase made up of everyday words is found far sooner than its length suggests. A generated passphrase is different: its words are picked at random, and its strength is counted on the assumption that the attacker has the whole word list.

3. The usual changes

Then every word is tried again with the changes people make to meet password rules: a capital at the start, 0 for o and @ for a, a year or 123 at the end, an exclamation mark, a word spelled backwards, a keyboard run like qwerty. Cracking tools such as hashcat come with thousands of these rules, so P@ssw0rd2024! is barely harder to find than password.

4. Every combination: brute force

A random password leaves only one option: trying every possible combination of characters. Each extra character multiplies the work, by 26 for a lowercase letter and by about 95 when it could be any character on the keyboard. That is why length beats cleverness, and why a random 16-character password is out of reach when a random 8-character one is not.

How This Checker Rates a Password

It reads your password the way a cracking tool would. It looks it up among the million most-leaked passwords, then looks for words in many languages, names, dates, keyboard runs, look-alike swaps, and words from our own passphrase lists. It finds the cheapest way to guess the password, counts how many guesses that takes, and shows how long 1,000 graphics cards would need. That is why five everyday words can rate Weak while 16 random characters rate Very strong.

How Long Would It Take to Crack Your Password?

Most checkers show a time without saying what it assumes, which is why the same password can get a few years on one site and billions of years on another. The answer depends on who is guessing and how the password was stored, so here is exactly what ours assumes.

The case we plan for. A stolen copy of a website’s passwords, guessed with nothing to slow it down. We assume the passwords were stored in a fast form, such as an MD5, SHA-1, or NTLM hash (the form Windows keeps sign-in passwords in), which many systems still use. The time shown is the average: on average a password is found halfway through every possibility.

Who is guessing. Pick 1 GPU, 1,000 GPUs, or 100,000 GPUs under Time to crack to change the attacker. Your result always uses 1,000.

One gaming PC. A high-end gaming graphics card, or GPU, checks about 200 billion guesses a second against a very fast hash such as MD5 or NTLM, going by published hashcat benchmarks. At that speed a random 8-character password that can use any keyboard character lasts hours, and a random 16-character one lasts longer than the universe has existed.

An organized group or a botnet: your result. A thousand graphics cards guess about 200 trillion times a second. They can be rented, so a well-funded attacker does not need to own them. It is a harsh, hacker-scale benchmark rather than a typical attack, and it is the attacker your result is measured against (see what the result means).

A tech giant or a government. 100,000 graphics cards show what a far bigger effort could do. No government publishes its figures, so this shows scale only and never sets the result.

Why a day, 10 years, and 100 years. The work is the same whoever does it: what 1,000 graphics cards do in a day, one card does in about three years. Renting 1,000 cards costs roughly ten to thirty thousand dollars a day, so a password that falls within a day is cheap to crack. Holding out for 10 years would cost tens of millions, and for 100 years hundreds of millions. Strong starts at 10 years, not one, because a stolen password list can be attacked for years, on hardware that keeps getting faster.

What the time leaves out. A site that stores passwords properly, with a slow hash such as bcrypt, can make every guess millions of times slower (Argon2 is built for the same purpose), and a sign-in page blocks repeated tries. But you can’t see how a site stores your password, and none of that helps if the password is phished, caught by malware, or reused on a site that leaks it. So we show the worst case worth planning for.

Wi-Fi. Every time a phone or laptop joins your network, it sends a scrambled proof of the password, called the handshake, and some routers give out a similar code without anyone joining. Someone within range can record it, take it away, and test guesses against it on their own computers, with no limit on tries and without your router ever knowing. WPA2 runs every guess through thousands of steps, so one graphics card manages about 3 million guesses a second instead of hundreds of billions. Networks set to WPA3 only stop this kind of guessing; routers in WPA2/WPA3 mixed mode don’t.

Password managers and encrypted drives. A copied password manager vault, or a drive locked by a password alone (a BitLocker USB drive, VeraCrypt), allows around a hundred thousand guesses a second per graphics card, or fewer. That speed matches older settings; current password managers and drives are often 10 to 100 times slower to guess. Laptops that keep the drive key in a security chip block this kind of guessing from a copied drive.

Encrypted files. A password-protected file is only as slow as its format. 7-Zip, RAR, and recent Word and Excel files are slow to guess, like a password manager. An old-style ZIP file, or a Word or PDF file made by old software, can be far faster, closer to a hacked website.

Phones and laptops. Phones and most new laptops lock or slow down after a few wrong tries, so fast guessing is impossible there. The exception is a computer without drive encryption (BitLocker on Windows, FileVault on a Mac): its sign-in password can be copied off the disk and guessed offline, and its files can be read without the password at all.

Wherever a password is used, a weak one is the first kind to fall. Slow guessing only makes it take longer.

Sources: Graphics card benchmarks (hashcat), zxcvbn (USENIX Security), NIST password guidelines, NIST on quantum computing, PassGAN (research paper). Word lists from FrequencyWords by Hermit Dave, CC BY-SA 4.0.

Is It Safe to Type Your Password Here?

Yes. The strength check runs in your browser and your password is not sent; once the page has loaded, it even works without a connection. Only the full breach check uses the internet, and it never sends the password. Nothing is sent as you type each character. When you stop typing or paste a password, this happens once, using the password password as the example:

  1. You type the passwordpasswordstays on your device
  2. Your browser turns it into a SHA-1 fingerprint5BAA61E4C9B93F3F0682250B6CF8331B7EE68FD8stays on your device
  3. Only the first 5 characters go to Have I Been Pwned5BAA6sent
  4. It sends back every leaked fingerprint that starts the same way, about 2,000, mixed with dummiesyour browser looks for the match

The service never sees the password or its full fingerprint, so it can’t tell which of those 2,000 you checked.

  • The quick leak check is on your device. The page downloads, once, a compact list of fingerprints of the million most-leaked passwords, so the check as you type needs no request at all. A false match is about one in a million.
  • One request per password, never one per keystroke. The check waits until you stop typing. A password already on the on-device list sends nothing at all, and if you change a password only slightly (a typo fixed, a few characters added), it is checked at most twice, because comparing near-identical checks could give it away. Some checkers send a lookup on every keystroke, which together give the password away one character at a time.
  • Nothing is kept. The password is never stored, never put in the address bar, and gone when you leave the page.
  • Check it yourself. Open the Network tab of your browser’s developer tools and type: nothing goes out until you stop. Then you will see one request, to api.pwnedpasswords.com, ending in the 5 characters shown in this section. Or disconnect from the internet: everything except the full breach check keeps working.

Leaked-password data: Pwned Passwords by Have I Been Pwned, used under the Creative Commons Attribution 4.0 license.

Will Your Password Still Be Safe in the Future?

Faster Graphics Cards

Each new generation of graphics cards guesses faster, and more of them can be rented. If attackers become ten times faster, every time on this page drops tenfold. Two more random characters make a password thousands of times harder to guess, which covers many years of progress. A random password of 16 characters or more has room to spare, and so does a passphrase of eight or more words picked at random by a generator.

AI Password Guessers

AI password guessers such as PassGAN learn from leaked passwords how people choose them, then produce guesses that look human. That is one more way to find passwords people made up. It is no help against a random password, because a random password has no human habit to learn.

Quantum Computers

They threaten some of the encryption used to set up secure connections, not a long random password. In theory a quantum computer could search for a password in roughly the square root of the usual number of guesses, but every step would have to run one after another on a machine far beyond anything built today, and a slow hash makes each step costlier still. NIST, the US standards agency, says this quantum method is “quite likely” to give “little or no advantage” against AES encryption.

In practice, reuse, phishing, malware, and leaks are more common risks than someone cracking a strong random password, and no length protects against them: see what to do.

Frequently Asked Questions

Is it safe to type my real password into a password checker?

Only into one that checks on your device and lets you prove it. This one works out the strength in your browser, and its breach check sends only 5 characters of a scrambled fingerprint, once, after you stop typing. Watch it in your browser’s Network tab, or disconnect from the internet and see it keep working.

How do I check if my password has been leaked?

Type or paste it above. If it is one of the million most-leaked passwords, you see it straight away. When you stop typing, it checks every known breach as well. If it appears at all, stop using it.

Which password is the most hacked?

123456, found in data breaches over 210 million times, followed by 123456789, 12345678, password, and qwerty (figures from Have I Been Pwned). Anything a person can think of quickly is on the lists.

How long should a password be?

At least 16 random characters, or seven random words from a generator (eight for anything that must hold up for years); words you pick yourself don’t count as random. Twelve random characters rate Strong against the worst case on this page when they include symbols, and only Fair without them. NIST, a US standards agency, tells websites to accept passwords of at least 64 characters and to require at least 15 when the password is the only thing protecting the account.

Can you give me an example of a strong password?

Treat any password you have seen published as unsafe, because it is now on the lists. Even correcthorsebatterystaple from the xkcd comic has turned up in breaches thousands of times. A strong password is one made at random just for you: use the password generator or the passphrase generator.

Why does a long password still show as weak?

Because cracking tools don’t guess one letter at a time. They guess whole words, names, places, dates, and common passwords, then join them and add numbers. LiverpoolFootball1892 has 21 characters, but to a cracking tool it is two common words and a year, which one graphics card guesses in under a second. Length only helps when the characters are unpredictable, as in a random password or a passphrase of random words.

Where can I see which of my saved passwords are compromised?

Your password manager can check them all at once. On an iPhone or a Mac, open the Passwords app and look under Security. In Chrome, Google Password Manager has Password Checkup. NordPass, Bitwarden, 1Password, and most other password managers have a similar report.

Is Have I Been Pwned safe to use?

Yes, for passwords. Your browser sends only the first 5 characters of the password’s fingerprint and does the final comparison itself. This page checks passwords only; to see whether your email address was in a breach, use Have I Been Pwned’s own site.

Why do other checkers show longer times?

Every checker has to assume an attacker, and most don’t say which. Many assume a site that stores passwords in a deliberately slow way, which allows only thousands of guesses a second. Some use graphics card speeds from years ago, and some count only the length and miss patterns like words and dates. Some, like GRC’s Password Haystack, count only the length and the kinds of characters and say themselves that they are not strength meters, so a password like BlueOcean#7 gets years there and Weak here. We show a harsher case and say so: a stolen password list guessed at full speed by 1,000 graphics cards. Many well-run sites are much slower to crack than that, but you can’t tell which kind a site is, so treat our times as the worst case to plan for.

Why does the checker show a different strength from the Password and Passphrase tools?

The generators know exactly how a password was made: its length, its kinds of characters, its word list. The checker only sees the finished password. Usually that makes the checker stricter: it can see a word, a name, or a pattern you never meant, such as a word spelled backwards, and it errs on the safe side. Two settings work the other way. With Leave out look-alikes, the generator counts only the characters it was allowed to use, and the checker cannot tell which ones were left out, so it can show a little more strength. With a maximum length on a passphrase, the checker cannot tell which limit you chose. Both answers are honest; for a password you made up yourself, the checker’s is the one to go by.

What is password entropy?

How unpredictable a password is, counted in bits: each bit doubles the number of guesses an attacker needs, so 10 more bits mean about a thousand times more guessing.

What is the 8 4 rule for passwords?

An old rule: at least 8 characters, using all four kinds (lowercase, capitals, digits, and symbols). Even a random password like that falls in hours to one graphics card, so it is no longer enough. NIST now tells sites to require at least 15 characters when the password is the only thing protecting the account.