Ask five creators the best clip length and you'll get five confident numbers. Most of them are superstition. The platforms stopped enforcing tight caps, the feeds rank on watch behaviour, and the honest answer is that length is downstream of the moment you picked. This piece covers the real 2026 limits, the metric that actually decides distribution, and practical ranges by moment type.
There is no universal best length. The platforms all take clips up to 3 minutes or more in 2026, and the feeds reward completion and rewatches, not any particular duration. The reliable rule: a clip should be exactly as long as its moment, hook first, payoff last, nothing after the payoff. In practice that means most strong clips land between 15 and 60 seconds, with length chosen by the shape of the moment, not superstition.
The 2026 platform limits
First, the hard constraints, which are looser than most advice assumes:
- YouTube Shorts: up to 3 minutes (extended from 60 seconds in late 2024). Shorts are detected by aspect ratio and length, not a special upload flow.
- Instagram Reels: up to 3 minutes for standard uploads (extended in early 2025).
- TikTok: uploads well beyond short-form, up to 10 minutes and longer in some regions, with the feed still centred on short clips. Note TikTok's monetization programs generally favour videos over one minute.
So the ceiling almost never decides your length anymore. The feed does. A fuller platform comparison is in Shorts vs TikTok vs Reels.
Completion rate is the metric
All three feeds distribute on watch behaviour, and the signals they reward are versions of the same thing: did people watch to the end, and did any of them watch twice. That is why length advice framed as a magic number misses the point. A 20-second clip watched fully beats a 60-second clip abandoned at 30 seconds, but a gripping 90-second story watched through beats them both. Length is not the lever; the ratio of watched time to total time is.
Two structural consequences. The hook has to happen immediately, viewers decide in the first second or two, so the clip must open on the hook line itself, not the setup. And nothing survives after the payoff: every second past the natural end is completion rate bleeding away. Trim the clip to end on the punchline, even if it feels abrupt in the editor. It won't feel abrupt in the feed.
Length by moment type
The practical way to pick length is to match it to the shape of the moment:
- One-liners and reactions (roughly 8-20s): a single quotable line, a punchline, a reaction spike. Cut brutally tight, these are the most rewatchable shape.
- Exchanges and hot takes (roughly 20-45s): an opinion with one supporting beat, or a question and its answer. Long enough to land an idea, short enough to hold.
- Stories and explanations (roughly 45-90s): a setup-tension-payoff arc or a genuinely useful how-to. These earn their length only when every sentence advances the arc.
What all three share: the clip is self-contained. If the first line needs context from before the cut, no length fixes it. That selection judgment, finding the moment whose natural shape fits a feed, is most of the craft, and it is the part worth directing the AI on when you run an automated pass.
When longer wins
The 3-minute era changed the calculus at the top end. Longer clips win when the material genuinely sustains tension: multi-beat stories, escalating arguments, demonstrations with a visible end state. On TikTok specifically, the over-one-minute threshold interacts with monetization, which is why you see more 60-90 second cuts there in 2026. But longer clips lose harder when they miss: abandonment in the first half of a 3-minute clip suppresses it more than a tight 30-second cut that most people finish. If in doubt, cut the shorter version. You can always post the extended cut as a second clip, which is also the cheapest way to A/B your own material.
Putting it into practice
When we tuned ClipDocker's selection, the rule that survived testing was: let the moment set the window and enforce clean edges. The AI scores moments for hook and payoff, then cuts so the clip opens on the hook line and ends on the payoff sentence. By default it picks lengths across the ranges above ("auto"), or you can pin a range, say 30-59 seconds, when a specific platform strategy calls for it. Whether you use our tool or cut by hand, the checklist is the same: hook in the first two seconds, self-contained throughout, ends on the payoff, and not one second longer than the moment deserves.
See what lengths your video naturally produces: 60 free signup credits, paste a link, and compare what the AI picks per moment. Try it free →
Frequently asked questions
How long can a YouTube Short be in 2026?
Up to 3 minutes. YouTube extended Shorts from 60 seconds to 3 minutes in late 2024; anything vertical and under that length is treated as a Short. The extension changed strategy at the top end, but most high-performing Shorts still land well under a minute.
Is 30 seconds or 60 seconds better for TikTok?
Neither is inherently better. TikTok's feed rewards completion and rewatches, so the right length is whatever the moment sustains. One practical wrinkle: TikTok's monetization programs generally favour videos over one minute, which is why creators chasing payouts often cut 60-90 second versions.
Does clip length affect how the algorithm distributes a video?
Indirectly. The feeds distribute on watch behaviour, primarily completion rate and rewatches, and length changes how hard those are to earn. A short clip is easier to finish; a long clip that holds attention accumulates more total watch time. Length is a tool for maximizing completion, not a ranking factor by itself.
What clip length does ClipDocker cut by default?
On auto, ClipDocker lets each moment set its own window: the AI scores hooks and payoffs, opens the clip on the hook line and ends it on the payoff, which lands most clips between 15 and 90 seconds. You can also pin a fixed range, like 30-59 seconds, per project.