CampaignReel

Ad placement specs against what generators output

Hedra publishes ten minutes, driven by audio

The Character-3 model page states a maximum duration of ten minutes. Nothing this register reads on the placement side comes close, so the cell is accurate and inert at once, and what it actually describes is how long a performance can be sustained. As of 2026-09-22.

Where a ten minute figure comes fromLength here follows the supplied audio track rather than how long a generated scene holds together, which is why the ceiling is measured in minutes.Audio trackRuns as long as the speechSupplied by the buyerSets the output lengthCharacter modelAnimates to the trackFrame by frameHolds one performanceTen minute ceilingA file handling boundStated on the pageNothing here reaches itA figure about speech, not about scenesWhat each part contributes to the number
Fig. 1 Two products answer the same brief and their length figures are not the same quantity.
The ceiling, and what governs it. Recorded 2026-09-22.
FigureValueWhat sets it
Max duration10mHow long the supplied audio runs
Longest ad slot in this register60 secondsA platform help page
HeadroomNine minutes on any slot recorded hereNot a usable quantity

Inclusion rule. The duration figure this model page states, set against the longest placement ceiling this register carries. Figures for other Hedra models are not recorded on this sheet. Order. The published figure first, then what it is being read against.

1A ceiling that nothing reaches is still worth printing

Ten minutes answers no question a media buyer has. Its value in the register is comparative: it marks this row as belonging to a different product category from the scene generators, whose longest single runs are measured in tens of seconds.

Reading it as headroom is the error the cell invites. Headroom implies the constraint has moved somewhere else, and here it has: to whether a speaking performance is the right form for the advertisement at all, which no spec sheet settles.

2Length governed by audio is length governed by a script

Where a model animates a supplied audio track, the output runs as long as the track. The ceiling is then a statement about file handling rather than about how long a coherent scene can be held together, which is what the same number means on a prompt-driven row.

That is why this register refuses to rank the column. Ten minutes here and thirty seconds on a scene generator are not the same quantity measured twice, and a table that sorted them into an order would be publishing a category error as a finding.

3What the resolution ladder adds to the reading

The same page publishes 540p, 720p and 1080p, with no condition tying any of them to the length. A ten minute piece at 1080p is available in the terms the page states, which is unusual: two other rows here couple their longest run to a resolution.

For a sixty second non-skippable that combination is the whole appeal. One request, one continuous performance, no joins to match and no tier to give up, against a format whose impression billing charges for every second including the ones a seam would spoil.

4Where this cell sits in the column

Duration, generator by generatorWhat each generator prints as a length, Hedra marked. Nothing on this axis comes from a surface. Recorded 2026-09-22.Duration, generator by generatorDurationAmazon Nova Reel — Duration: 6 s increments, up to two minutesAmazon Nova ReelCogVideoX — Duration: 6 s, at 49 framesCogVideoXGoogle Veo — Duration: 4, 6 or 8 sGoogle VeoGrok Imagine — Duration: 1 to 15 sGrok ImagineHedra — Duration: 10 minutesHedraHeyGen — Duration: 30 minutes per sceneHeyGenKling AI — Duration: Not statedKling AILTX-Video — Duration: Frames, never secondsLTX-VideoMiniMax — Duration: 4 to 15 sMiniMaxPika — Duration: 5 s and 10 s, 20-25 s on PikaframesPikaPixVerse — Duration: 1 to 15 s on version 6PixVerseRunway — Duration: 5 to 15 s on Gen-4.5RunwaySceneMixer — Duration: Thirty seconds on the Wan tiers, fifteen on H3SceneMixerVidu — Duration: Not statedViduWan 3.0 — Duration: 2 to 30 sWan 3.0
Fig. 2 What each generator prints as a length, Hedra marked. Nothing on this axis comes from a surface. Recorded 2026-09-22.
Every generator in the register on the duration column, with Hedra marked. Recorded 2026-09-22.
GeneratorDuration
Amazon Nova Reel6 s increments, up to two minutes
CogVideoX6 s, at 49 frames
Google Veo4, 6 or 8 s
Grok Imagine1 to 15 s
Hedra10 minutes
HeyGen30 minutes per scene
Kling AINot stated
LTX-VideoFrames, never seconds
MiniMax4 to 15 s
Pika5 s and 10 s, 20-25 s on Pikaframes
PixVerse1 to 15 s on version 6
Runway5 to 15 s on Gen-4.5
SceneMixerThirty seconds on the Wan tiers, fifteen on H3
ViduNot stated
Wan 3.02 to 30 s

Inclusion rule. Every generator with a row in this register, carrying whatever its own page states as a length. A page that states none keeps its row and says so. Order. Alphabetical by generator name.

5Sources

Read from the Hedra Character-3 model page, checked 2026-09-22. The column itself is described on duration, and the rule that keeps surfaces and generators apart is on method. Next along this column: HeyGen, Kling AI.