Short form against long form

16 Sep 2026, 16:28 · counted on both datasets, with the sampling bias stated
Why this matters for the tool, before any number. Vidrush makes long-form and it says so itself - no vertical video. So short form is not a thing the tool can produce today. The question this page answers is a different one: is the money on the other side of the fence? If shorts beat long form badly enough, the product is aimed at the smaller half.

The two datasets

GTA 5, the whole archive

thirteen years of uploads
Short formLong form
videos4,60810,820
median views3.3M2.5M
mean views8.4M5.6M
total views38.8B60.1B
Biased sample - see the warning below.

GTA 6, the fortnight after the Extended Look

swept with no view bar, so it is fair
Short formLong form
videos1,4251,094
median views19k35k
mean views147k236k
total views210M258M
Long form wins on median, mean and total.
Read the left card with care - it is not a fair fight. The GTA 5 archive was collected with two different bars: short form only if it passed 2,000,000 views, long form if it passed 500,000. So the shorts in it are the top of the top and the long videos are not. Their medians look similar only because the shorts had a four times higher entry fee. That number cannot be used to compare the two formats, and it is here so nobody quotes it later as if it could be.

The right card is the honest one. The GTA 6 fortnight was swept without a view bar, so both sides were counted the same way. There, long form wins on every measure: a higher median, a higher mean, and more total views from 336 fewer videos.

What follows for the product

Long form is not the smaller half254Mtotal views for long form in the GTA 6 fortnight, against 210M for short form
A short has no scriptthe tool turns a script into footage; a 20-second clip has nothing to turn
Vertical is out anywayVidrush states it does not do vertical video, so shorts are not a switch we can flip
Where shorts would payas cut-downs of a finished long video, which is an export setting, not a second product

The length bands inside long form

Only the videos in the corpus - the ones the tool could rebuild. Longer does not cost you views.
8 to 15 min751k693 videos
15 to 30 min677k308 videos
30 to 60 min732k106 videos
over an hour679k73 videos