Sora

Sora is OpenAI's text-to-video model. It generates short clips from a written description, with options to extend, blend and re-cut scenes.

What it does well

• Prompt fidelity. Complex multi-subject descriptions are handled more reliably than in earlier video models.
• Coherent motion. Camera movement and subject consistency hold up across a short shot.
• Storyboard tooling. Scenes can be sequenced and remixed inside the same interface.
• Resolution and duration control. Output settings are exposed rather than fixed.

Limitations

Physics still break under close inspection, long continuous shots drift, and generation cost and queue times are significant. Availability depends on plan and region.

Best for

Concept work, previsualisation and short social clips rather than finished broadcast footage.