Minesweeper world records: the players, the times, the rules

The three difficulties everyone races

Three board sizes are treated as the canonical record categories, unchanged since the Windows port shipped in 1990. Every world record you'll see quoted lives in one of these three boxes.

• Beginner — 9×9, 10 mines. 12% mine density. Records are sub-1 second.
• Intermediate — 16×16, 40 mines. 15.6% density. Records are sub-8 seconds.
• Expert — 30×16, 99 mines. 20.6% density. Records sit in the high-20-second range.

Each category splits into two sub-categories: with flags allowed, and no-flag (NF). NF runs are faster on average because every flag costs a click — top players skip flagging entirely on beginner and intermediate.

The names you'll see at the top

Kamil Murański

Polish player; one of the longest-running holders of the expert record. Murański popularized many of the modern speed techniques — in particular, ruthless opening selection and a near-zero flagging habit on intermediate. His expert record at one point stood for years without challenge.

Dion Tiu

Indonesian player who pushed the expert record into territory most analysts considered theoretically impossible — sub-30 NF on Arbiter (the standard verification client). Tiu's record run is one of the most-watched minesweeper videos on YouTube.

Andrey Ognev

Russian player known for his consistency across difficulties. Ognev's name shows up at the top of all three boards at various times. His intermediate runs in particular set the modern sub-8-second baseline.

The rest of the top 20

Records below the absolute world tier are still genuinely elite — sub-35 expert and sub-9 intermediate are achievable by maybe a few hundred players globally. Most of them hang around the Authoritative Minesweeper Forum and the Minesweeper Game World community.

How records are verified

Anyone can claim a fast time. The community only counts a record when three pieces of evidence line up:

1. A full-screen video recording of the entire game from the moment the timer starts to the moment the last cell is revealed. Cut footage is rejected outright.
2. The original replay file from a supported client (Arbiter, Viennasweeper, or — increasingly — public web clients that export a deterministic seed plus an input trace). The replay lets reviewers step through the run cell-by-cell.
3. A 3BV/s figure consistent with the recorded inputs. 3BV is the minimum number of clicks needed to clear a given board; 3BV/s is your effective click rate. A claimed time that implies inhuman clicks-per-second triggers a re-review.

Records on non-canonical clients (random web pages, mobile apps without replay export) are tracked as "informal" but not used in official ranking. That's why we don't claim the Minesweeper Battle daily leaderboard times are world records — they share a seed, which is great for fair competition but not how the WR community validates runs. If you want to chase a real WR, do it on Arbiter.

What it would take to break a record

Pure click speed isn't the bottleneck for any modern WR holder. Every top player can sustain 7-10 clicks per second on a hot streak. The real bottlenecks are different:

• Pattern recognition speed. A record run never pauses to think. The deductions are pre-loaded — you see the 1-1 pattern or 1-2 pattern the instant it appears and your hands move before your conscious mind names it.
• Chord proficiency. WR runs use the chord move on essentially every satisfiable number. Players who manually reveal each neighbour are leaving 3-5 seconds on the table per expert board.
• Restart discipline. Top players abandon a run within the first three seconds if the opening flood-fill is small. The single most underrated record-chasing skill is mashing the restart key fast enough that nine bad openings cost less time than fighting through one of them.
• Lucky openings. No amount of skill produces a 25-second expert run without a generous initial flood-fill. Even the WR holders are seed-fishing — they just do it faster.

How to start tracking your own progress

Don't aim at the WR. Aim at the next quartile up from where you are. A useful self-test ladder:

1. Under 60 seconds expert — top ~1% of casual players.
2. Under 100 seconds expert — comfortably skilled. See the expert strategy guide.
3. Under 150 seconds expert — solid intermediate ability scaled up.
4. Anything you finish — you've already beaten the median internet user.

Use a daily metric, not a one-off PR. Records are about consistency over thousands of attempts — track your median time over the last 50 wins, not your fastest ever.

The daily minesweeper challenge gives you a shared seed every 24 hours, so you can directly compare your time against players who solved the exact same board. It's the closest thing to a world-record framework you'll get without setting up Arbiter.

Frequently asked questions

What is the minesweeper world record?

On expert the current public records sit in the high-20-second range. Intermediate WRs are sub-8 seconds; beginner WRs are sub-1 second. The exact figures move as new runs get verified, so the Authoritative Minesweeper Forum is the canonical reference.

Who is the fastest minesweeper player ever?

It depends which difficulty and which year. Kamil Murański, Dion Tiu, and Andrey Ognev have all held top spots; the community treats the modern era as a multi-holder rotation rather than a single GOAT.

Can I set a world record on Minesweeper Battle?

Not formally. We don't export Arbiter-compatible replay files, and the daily challenge uses shared seeds (which is great for comparison and terrible for fairness in WR terms). Our leaderboards are competitive but informal.

Is the first click safe in a record run?

Yes — every modern minesweeper client (including the ones used for WR validation) guarantees the first click is not a mine. WR runs assume this; the run starts with a click that triggers the largest possible flood-fill the seed allows.