28 s
median length of a viral TikTok
n = 1,696 viral TikToks
45 s
median length of a TikTok from the same searches that did not go viral
n = 306 TikToks
1.8x
as many breakouts as expected for Reels under 15 seconds (3 times the creator’s median or more)
95% interval 1.16 to 2.64; 25 breakouts in 121 Reels
19 s
median length of the most viral grade, Exceptional
n = 928 TikToks
Viral TikToks are shorter than the posts beside them
We compared the viral TikToks in our library with the posts the same searches found that did not pass their creator’s follower count. The viral ones had a median length of 28 s, with a middle half of 13 s to 64 s. The others had a median of 45 s (22 s to 88 s). The most viral grade, Exceptional, was shorter still: a median of 19 s, with 38% under 15 seconds.
An older, separate sample from our first viral finder, collected with a different search sort months earlier, shows the same gap: a median of 36 s for viral TikToks against 61 s for the rest (2,817 and 388 posts).
- Viral (n = 1,696)
- Not viral (n = 306)
- Under 15 s29% Viral (n = 1,696)17% Not viral (n = 306)
- 15 to 30 s22% Viral (n = 1,696)18% Not viral (n = 306)
- 30 to 60 s18% Viral (n = 1,696)26% Not viral (n = 306)
- 60 to 90 s14% Viral (n = 1,696)14% Not viral (n = 306)
- 90 s to 3 min11% Viral (n = 1,696)18% Not viral (n = 306)
- 3 min or more5.1% Viral (n = 1,696)6.2% Not viral (n = 306)
Search library, same searches: 1,696 viral and 306 not viral TikToks.
| Group | Posts | 25th percentile | Median | 75th percentile | 90th percentile |
|---|---|---|---|---|---|
| Viral | 1,696 | 13 s | 28 s | 64 s | 140 s |
| Not viral, same searches | 306 | 22 s | 45 s | 88 s | 156 s |
| Exceptional grade only | 928 | 10 s | 19 s | 51 s | 94 s |
| Older sample, viral | 2,817 | 15 s | 36 s | 75 s | 159 s |
| Older sample, not viral | 388 | 22 s | 61 s | 121 s | 188 s |
| Group | Under 15 s | 15 to 30 s | 30 to 60 s | 60 to 90 s | 90 s to 3 min | 3 min or more |
|---|---|---|---|---|---|---|
| Viral | 29% | 22% | 18% | 14% | 11% | 5.1% |
| Not viral, same searches | 17% | 18% | 26% | 14% | 18% | 6.2% |
| Exceptional grade only | 38% | 25% | 16% | 11% | 7.2% | 3.6% |
| Older sample, viral | 24% | 22% | 16% | 18% | 13% | 7.4% |
| Older sample, not viral | 16% | 16% | 16% | 17% | 23% | 11% |
Rows may not add up to exactly 100% because each share is rounded on its own.
Read this as a description, not a rule. By our definition, a post with many views is not viral only when its account is bigger still, so the non-viral group leans toward larger accounts, and part of the gap may be about who makes longer videos. That is why we also compared creators with themselves, below.
How long can a TikTok be?
TikTok’s own help page on camera tools gives two ceilings: “Videos you record in TikTok can be up to 10 minutes long.” and “Videos you upload in TikTok can be up to 60 minutes long.” It adds: “If you select a sound before you record or upload your video, the video length will be determined by the sound length.” (TikTok Support, “Camera tools”, read November 14, 2026). On a computer, the whole file goes up: “When you upload a video on your web browser, your entire video will be uploaded. To change the length of the video, upload your video in the TikTok app.” (TikTok Support, “Making a post”, read the same day).
Neither page gives a minimum length or a recommended one, and neither says that length changes how many people see a video. So the limit tells you what TikTok accepts, not what to aim for.
How long a TikTok should be is the question this study answers with data. The viral TikToks in our library ran a median of 28 s, with a middle half of 13 s to 64 s, far under either limit. Long ones went viral too: 11% of the viral TikToks ran 90 seconds to 3 minutes, and 5.1% ran 3 minutes or more. The comparison with the posts that did not go viral, and why it is a description rather than a rule, is in the first section of this page.
Within the same creators, short Reels broke out more often
Length and account size travel together, so we also compared each creator with themselves. For Instagram creators our users track, we marked a Reel as a breakout when it reached at least 3 times that creator’s median Reel views, counting Reels at least 3 days old from the last 365 days, leaving out paid and pinned posts, and only creators with at least 10 such Reels. That gave 1,437 Reels from 39 creators, 214 of them breakouts. Then, for each length band, we compared the breakouts we saw with the number expected from those same creators’ usual breakout rates.
Reels under 15 seconds broke out 1.8 times as often as expected: 25 breakouts where 14 were expected, in 121 Reels from 25 creators (95% interval 1.16 to 2.64). No other band differed clearly from what was expected; Reels of 90 seconds to 3 minutes came in at 0.83 (0.53 to 1.23).
Two limits matter here. The result rests on only 25 breakout Reels, so the true effect could be much smaller (the interval starts at 1.16). And a Reel that loops counts each replay as a view, which favours short Reels without meaning more people watched them.
- Under 15 s1.79 (1.16 to 2.64)
- 15 to 30 s1.16 (0.79 to 1.63)
- 30 to 60 s0.89 (0.71 to 1.10)
- 60 to 90 s1.00 (0.72 to 1.35)
- 90 s to 3 min0.83 (0.53 to 1.23)
Dot: observed over expected. Bar: 95% interval. Line: 1, no difference. 39 tracked Instagram creators; Reels of 3 minutes or more left out (too few).
| Reel length | Reels | Creators | Breakouts | Breakout rate | Expected | Observed over expected | 95% interval, low | 95% interval, high |
|---|---|---|---|---|---|---|---|---|
| Under 15 s | 121 | 25 | 25 | 21% | 14 | 1.79 | 1.16 | 2.64 |
| 15 to 30 s | 165 | 36 | 32 | 19% | 28 | 1.16 | 0.79 | 1.63 |
| 30 to 60 s | 549 | 39 | 87 | 16% | 98 | 0.89 | 0.71 | 1.10 |
| 60 to 90 s | 335 | 34 | 43 | 13% | 43 | 1.00 | 0.72 | 1.35 |
| 90 s to 3 min | 249 | 30 | 24 | 9.6% | 29 | 0.83 | 0.53 | 1.23 |
A breakout is a Reel with at least 3 times its creator’s median Reel views. Expected = the breakouts these Reels would have had at each creator’s own usual rate. Reels of 3 minutes or more are left out: too few for our thresholds.
Captions and engagement, described
Viral TikToks had shorter captions: a median of 119 characters against 210 for the posts that did not go viral. Their first sentence, taken from the transcript when we had one and otherwise from the caption, ran 8 words against 12. Hashtags did not differ much: a median of 5 in both groups, though 23% of the non-viral posts had none, against 11% of the viral ones.
Viral TikToks also had more likes and shares per view: a median like rate of 9.7% against 5.1%, and a share rate of 0.58% against 0.22%. Comments per view were lower (0.05% against 0.08%).
These numbers describe the two groups. They do not show that a short caption or a high like rate made a post travel: the groups differ in account size, and posts shown to many people collect engagement in their own way.
The TikTok engagement rate study takes these posts further: engagement per view by view count and account size, how the first sentence opens, caption and hashtag bands, and the Reels of creators we track.
| Measure | Viral | Not viral |
|---|---|---|
| posts | 1,696 | 306 |
| caption length, median characters | 119 | 210 |
| words in the first sentence, median | 8 | 12 |
| hashtags, median | 5 | 5 |
| share with no hashtag | 11% | 23% |
| share with 6 or more hashtags | 2.4% | 5.2% |
| likes per view, median | 9.7% | 5.1% |
| comments per view, median | 0.05% | 0.08% |
| shares per view, median | 0.58% | 0.22% |
The first sentence comes from the transcript when we had one, otherwise from the caption, so treat its length as rough.
What this means for your videos
- Short is common among viral TikToks, not required: 11% of the viral ones ran 90 seconds to 3 minutes and 5.1% ran longer.
- TikTok’s own growth advice leans the other way, toward longer videos. Our guide on how to go viral on TikTok quotes it next to this study and our other three.
- Check each platform’s ceiling before you film: TikTok’s is above, and our guides to how long an Instagram Reel can be and how long YouTube Shorts can be quote the limits from each platform’s own help pages.
- If an idea fits in under 30 seconds, try that cut first, and make every line earn its place. The Reel script generator writes a tight script you can time.
- Study how short viral posts in your niche are paced: the TikTok transcript generator gives you the words of any public TikTok, so you can see how fast the point arrives.
- To find those posts, run the free niche check, which ranks recent TikToks and Reels by views against followers. For the view numbers behind “viral”, see how many views is viral.
Limits of this data
- Loops count as views. When a short Reel replays, each replay adds to its view count, so a short Reel can collect more views without more people watching. That favours short Reels in the within-creator check.
- The Reels result is small. It rests on 25 breakout Reels under 15 seconds from 25 creators. The interval runs from 1.16 to 2.64, so the true effect could be much smaller than 1.8 times.
- Viral and not viral are not matched on size. A post with many views counts as not viral only when its account is bigger still, so the non-viral group leans toward larger accounts. Part of the TikTok length gap may be about who makes long videos.
- Search results, not a random sample. Every TikTok here was found by platform search, which shows big posts first, and most of the posts it surfaces already pass the viral line.
- The niche mix follows our users. Business, marketing, fitness, real estate and money topics are over-represented, and most captions are in English. TikToks were mostly published in October 2026.
- Tracked creators are not a random sample. The Reels check uses creators our users chose to follow, mostly large accounts.
- Captions and engagement are descriptive. They describe the two groups and say nothing about cause.
The full method, sampling and definitions are on the methodology page.
See what is viral in your niche
These numbers describe many niches at once. To see the posts beating their creator’s following in yours, type your niche: the free niche check ranks recent TikToks and Reels by views against followers, the same test this research uses. It needs no account.
Frequently asked questions
How long should a TikTok be to go viral?
No length guarantees it, but viral TikToks in our data were short: a median of 28 s against 45 s for the posts the same searches found that did not go viral (1,696 and 306 posts). The most viral grade, Exceptional, had a median of 19 s.
How long can a TikTok be?
TikTok’s help page on camera tools says: “Videos you record in TikTok can be up to 10 minutes long.” and “Videos you upload in TikTok can be up to 60 minutes long.” (read November 14, 2026). It gives no minimum and no recommended length. Viral TikToks in our data ran far shorter than either limit: a median of 28 s.
Do shorter Reels get more views?
Within the same creators, Reels under 15 seconds reached 3 times their creator's median views 1.8 times as often as expected (95% interval 1.16 to 2.64). That rests on 25 breakout Reels, and replays count as views, which helps a short Reel that loops.
Can long videos go viral?
Yes. 11% of the viral TikToks in our library ran 90 seconds to 3 minutes, and 5.1% ran 3 minutes or more.
Does caption length matter on TikTok?
Viral TikToks had shorter captions in our data, a median of 119 characters against 210 for the rest, but that describes the two groups. It does not show that a shorter caption helps a post travel.
How many hashtags do viral TikToks use?
A median of 5, the same as the posts from the same searches that did not go viral (5). 11% of viral TikToks used none, against 23% of the others.
How to cite this
CreatorsCompanion, “Viral video length: how long viral TikToks are, and which Reels break out,” published , https://www.creatorscompanion.app/research/viral-video-length
You are welcome to quote these numbers with a link to this page. Every figure on the page, with the file and field it comes from, is in the CSV download. Spotted an error? Email support@creatorscompanion.app and we will correct the page and date the change.