Password Cracking Time in 2026: What AI Can and Cannot Tell You
Understand online login attacks, offline password cracking, and the assumptions behind 8-, 12-, and 16-character estimates before trusting an AI cracking chart.

There is no universal answer to how long AI needs to crack an eight-character password. An attacker guessing through a website faces different limits from someone testing a stolen password database on their own equipment. A useful estimate must describe the attack before giving a time.
If you came looking for a 2026 cracking chart, start with this distinction. A dramatic number without the hashing method, guessing rate, and password-selection assumptions cannot tell you whether your account will be compromised.
Video transcript
How fast can a password be cracked?
It depends on the attack.
A login page can limit guesses.
A stolen database is tested offline.
Predictable passwords are tried early:
names, dates, keyboard patterns.
Long and random is much harder to guess.
Reused it? One leak can open the rest.
Give every account its own password.
Leaked? Change it wherever you used it.
And turn on two-step verification.
Need to share a password?
Skip the chat: use a secret memo.
Send the password another way.
Share the link. Keep the secret.
Online login attempts and offline cracking
In an online attack, guesses go through the service's login system. Rate limits, account protections, and multifactor authentication affect what an attacker can do. Attackers may also try passwords stolen from another site, rather than exhaustively searching every combination.
In an offline attack, an attacker already has password hashes and tests guesses locally. The service's login rate limit does not govern those tests. The hashing algorithm, its cost settings, and available hardware matter. NIST's authentication guidance distinguishes these threats and recommends protections for both.
Neither situation is the same as stealing a signed-in session or persuading someone to enter a code on a phishing page. A long password helps with guessing; it does not solve every route into an account.
What an 8-, 12-, or 16-character comparison actually means
| Password description | What you can reasonably infer |
|---|---|
| Eight characters, chosen by a person | Common words and predictable substitutions may be tried early; no fixed cracking time follows from length alone. |
| Twelve characters, chosen by a person | More length helps, but a familiar quotation, name, or reused password can remain predictable. |
| Sixteen independently random characters | A much larger search space than a shorter password using the same alphabet; still no universal time guarantee. |
| Several independently random words | Can offer useful length and memorability; a famous example copied from an article is no longer a secret choice. |
These are comparisons, not measured benchmarks. A PIN protected by a device's attempt limits cannot be assigned the same cracking time as an unthrottled numeric secret.
How to read a cracking-time chart
Before trusting a number, look for:
- Attack model: online guessing or offline testing of stolen hashes.
- Storage details: hashing algorithm and cost settings, including whether salts are used.
- Equipment and rate: measured guesses per second on specified hardware.
- Password generation: truly random choices or a model of human patterns.
- Meaning of the result: average search time, worst case, or success against a particular dataset.
Under a simplified uniform model with N equally likely candidates, no repeated guesses, and a constant rate of R guesses per second, the expected time to find the password is (N + 1) / (2 × R), approximately N / (2 × R) when N is large. This does not describe how early a dictionary attack finds a popular password. Changing the assumed rate can radically change a chart without changing your password.
Where AI fits
AI can help rank likely guesses based on patterns in training data. Conventional dictionaries and rules also prioritize guesses. Neither creates a single speed that applies to every service or every password hash, and a model's success on one leaked dataset is not a promise about your account.
Names with dates, keyboard walks, and predictable letter replacements remain poor choices. Avoid using any example printed here as your real password.
Practical protection
Use a password manager to generate a long, unique password for each account. If you must memorize one, choose words independently at random rather than assembling a personal sentence. Enable a passkey where supported or appropriate multifactor authentication. Keep recovery options secure too.
If a password is exposed, change it on the affected service and anywhere it was reused. Review sessions and recovery settings. Do not paste your real password into an unfamiliar cracking-time checker.
For necessary handoffs, send only the required information to the intended person. A password-protected memo can keep a secret out of a regular chat thread, but a recipient can still copy it; expiry cannot recall a copy already made.
Frequently asked questions
Is every eight-character password cracked in minutes?
No. That statement omits the attack, storage method, hardware, and guessing controls. Short, predictable passwords are still a reason to improve your setup.
Does using sixteen characters guarantee safety?
No. Randomness, uniqueness, phishing resistance, and account recovery also matter. Sixteen predictable or previously leaked characters are not equivalent to sixteen random ones.
Should I change a strong password just because a new year starts?
A calendar change is not evidence of compromise. Change exposed or reused passwords, and improve weak ones; avoid replacing a strong unique password with a predictable yearly variation.
Read next: creating strong passwords, checking for leaked passwords, and passkeys versus passwords.
Create link
Create linkKeywords
You might also like
Best Password Managers Compared — 2026 Guide
Compare the best password manager approaches: browser built-in, 1Password, Bitwarden, LastPass, and more. Pros, cons, pricing, and what to look for when choosing one.
Passkeys vs Passwords — What You Need to Know in 2026
Understand how passkeys work, how they compare to traditional passwords, and whether you should start using them. Covers FIDO2, WebAuthn, pros and cons, and current adoption status.
How to Check If Your Password Has Been Leaked
Learn how to find out if your passwords were exposed in a data breach. Step-by-step guide to using Have I Been Pwned, Google Password Checkup, and what to do if your credentials are compromised.