Key takeaways
- Jay12310 finished a random-seed glitchless run in 6:56 on 5 February 2026, the second-fastest time ever submitted in the category.
- The world record is still lowk3y_’s 6:50 from June 2025 – the top of this leaderboard has not moved in eight months.
- Three of the five fastest runs in Minecraft history are now under seven minutes, which was not true two years ago.
There is a specific kind of run that does not break a record and still matters. Jay12310 submitted one on 5 February 2026: a random-seed glitchless completion in 6:56, verified on the 1.16-1.19 leaderboard, which slots into second place behind a world record that has now stood since the middle of last year.
Six seconds is the gap. In a category where the whole run lasts under seven minutes and every phase is a negotiation with generation luck, six seconds is roughly one good fortress spawn.
Where it sits
The top of the random-seed board is tight and getting tighter. lowk3y_’s 6:50 from 11 June 2025 remains the record. Jay12310’s 6:56 is now second. drip120’s 7:01, which was the benchmark for most of 2024, has been pushed to third.
What that table shows is a category compressing rather than advancing. Nobody has found a new trick that removes a minute; runners are shaving seconds off a route that is close to solved, and the result is a cluster of times within fifteen seconds of each other rather than a staircase.
What a 6:56 requires
Everything, in the right order, with no pauses to think. A village or ruined portal inside the first ninety seconds for the opening kit. A nether portal built from a lava pool rather than a mined bucket of obsidian. A fortress close enough to the portal that travel time does not eat the run. Enough blaze rods on the first few kills. Enough pearls from bartering that a second trip is unnecessary.
None of those are skills in the usual sense – they are dice rolls the runner is fast enough to capitalise on. The skill sits underneath: knowing instantly whether the seed you have been handed is worth continuing with, and resetting without hesitation when it is not.
Who Jay12310 is
Not a household name outside the category, which is normal. Minecraft speedrunning’s top ten is full of runners with four-figure follower counts and four-digit reset totals, and the sport rewards the people willing to sit through the resets more than it rewards the people who stream well. A run like this is usually the visible tip of several hundred hours of attempts that ended in the first two minutes.
That is worth saying plainly because the category is often described as luck. It is luck in the same sense that poker is: the cards are random, and the people at the top are still the same names year after year.
Why the record is not falling
The honest answer is that the route has been optimised to the point where the remaining gains are mostly luck. Runners like Jay12310 are producing near-perfect executions regularly; what they are waiting for is a seed that hands them a fortress and a bastion close to the portal on the same attempt.
That is a different kind of waiting from the one the category had in 2021, when each new record came with a new technique. It also means the next record, whenever it lands, is likely to be a small step rather than a leap – unless something structural changes in how the run is routed.
What changed in the route
Two things have shifted since drip120 set 7:01 two years ago. The first is bastion routing – runners now treat the bastion as the primary pearl source rather than a fallback, which removes an entire enderman-hunting phase from the overworld. The second is the zero-cycle dragon kill, which ends the run without waiting for the dragon to perch.
Both are now standard rather than optional, which is precisely why the leaderboard has compressed. When everybody runs the same optimal route, the difference between second place and the record stops being knowledge and becomes generation.
The other ladder
Runs like this one used to be the entire competitive picture. They are not any more. The MCSR Ranked ladder, which reset into its tenth season at the start of January, measures something a leaderboard cannot: consistency against a live opponent on a shared seed.
Both matter, and they reward different players. A 6:56 says Jay12310 can produce a top-three run in the history of the game. A ranked rating says whether they can do it against someone sitting opposite them, repeatedly, on seeds neither of them chose. The runners who do both are rare, which is exactly why the scene keeps two scoreboards.
For now the number to watch is still 6:50. It has survived a Minecraft version change, an entire ranked season and a queue of runners who are demonstrably capable of beating it. The interesting question is not whether it falls but what the run that beats it looks like – and whether anyone will be able to call it skill rather than a good afternoon. Our overview of where competitive Minecraft stands goes into that split in more detail.