OpenAI · AI Video Model
Audio

Sora 2

OpenAI's Sora 2: realistic visuals with synchronized audio — and face uploads that put you or your cast in the scene.

Browse All Models

What is Sora 2?

Sora 2 brought OpenAI's video line into the audio era: visuals are generated together with synchronized dialogue, ambience and effects, so a finished clip sounds complete as well as looks complete. Its physical realism — weight, materials, believable failure — set the standard its peers are measured against.

Its most distinctive feature on DreamDance is face support: upload a face and the model can place that person into generated scenes with their identity intact. That single capability turns Sora 2 from a footage generator into a casting tool — put yourself, a presenter, or a recurring character into any scenario you can describe.

Key Capabilities

Synchronized audio

Dialogue, ambience and effects generated with the picture, in sync from the start.

Face uploads

Cast a real person into generated scenes with their likeness preserved.

Physical realism

Materials, momentum and cause-and-effect behave the way viewers expect.

Flagship lineage, faster

Most of Sora 2 Pro's quality at a noticeably lighter render cadence.

Prompt Examples

Copy any of these straight into the workspace.

Cameo scene
[Face upload: your photo] You sit at a rooftop bar at sunset and raise a glass to camera: "To new beginnings." City hum below, wind gust, warm golden light, medium close-up.
Physics-forward shot
A stack of ceramic plates slides off a counter and shatters on tile in slow motion, each shard catching kitchen light. Static camera, crisp impact audio, natural reverb.

What Sora 2 Is Best For

  • Realistic scenes with dialogue that must sound right
  • Cameos and personal content via face upload
  • Physics-sensitive action and everyday realism
  • Premium work that cannot justify Pro render times

200s

Est. Latency

OpenAI

Model Vendor

1080P HD

Max Quality

Sora 2 FAQ

How does the face upload work?

Upload a clear photo of the person in the Sora 2 workspace, then reference them in your prompt. The model preserves their likeness across the generated scene — a cameo tool, not a photo edit.

What is the difference between Sora 2 and Sora 2 Pro?

Same family, different render budgets. Pro adds headroom for the most complex scenes and the most authored looks; standard Sora 2 delivers the family's realism and audio at a friendlier cadence.

How long does a generation take?

On DreamDance, Sora 2 averages a bit over three minutes per generation.

How much does it cost?

Generation costs credits that vary by model, resolution and duration. DreamDance plans start at $19.99/month with 1,999 credits — see the pricing page for details.

Explore Related Models

View all

Ready to Create with Sora 2?

Every model is included in your plan and billed in credits — no setup, no API keys, generation starts in seconds.

See Pricing