video model

ByteDance logo

Seedance 2.5

ByteDance

Text-to-video with natively synchronised audio — one pass, not a dub.

Seedance 2.5 generates picture and sound together rather than generating video and dubbing it afterwards, which is why dialogue actually lands on the lips. It also accepts reference images, reference audio for voice cloning, and reference video for motion — enough control to produce on-brand footage rather than generic stock.

Use it on HumanLike
CapabilitiesNative synchronised audioReference imagesVoice cloning from reference audioReference video motion

Sample output

What it actually produces.

One clip is worth more than any benchmark row. Where we have not published a sample yet, the frame below holds the space it will occupy.

Demo clip

No sample of Seedance 2.5 uploaded yet.

Text-to-video clip with natively synchronised dialogue.

Variants

2 ways to run it.

Same family, different quality/cost rung. Picking the wrong one is the most common way teams overpay.

Pro

High
seedance-2.5-pro

Highest fidelity, longest clips.

rel. cost
92
quality

Standard

Balanced
seedance-2.5

The default quality/cost point.

rel. cost
86
quality

Fit

Two audiences, one model.

The same model is a different proposition depending on what you point it at.

64/100

Recruitment fit

Not for assessment — for the top of the funnel. Job ads and employer-brand clips that would otherwise never get produced.

Best for

  • Job ad video
  • Employer-brand social clips
  • Careers-page b-roll

Strengths

  • Synchronised audio means a job ad ships without a separate VO pass
  • Reference images keep footage on-brand
  • Cheap enough to make a clip per role rather than one per quarter

Watch out for

  • Generated people in recruitment marketing can misrepresent your actual workforce — disclose it
  • No place anywhere near candidate evaluation

The numbers

Measured, not asserted.

HumanLike evals

Run on our own harness.

Wired into HumanLike
Yes — /seedance-2-5Our eval

Compare

Worth putting side by side.