Generate videos using OpenAI video generation models

Last updated on Jan 16, 2026

Learn to use OpenAI video model Sora 2 to generate videos using text prompts in the Firefly video editor.

Firefly video editor includes the Sora 2 video generation models, which enable you to generate high-quality videos and add them to the timeline for editing. 

On the Firefly homepage, select Create from the left panel and then Edit video (beta).

Open a project or create a new one to access a timeline.

Select the Generation settings   icon from the right panel next to the video preview.

Enter a descriptive text prompt in the prompt box explaining what you want to generate.

In the Model dropdown menu under General settings, select Sora 2.

Select the aspect ratio for your video—Widescreen (16:9) or Vertical (9:16).

Select the desired Duration for the generated video from 4 seconds, 8 seconds, and 12 seconds.

If you want to add a first frame for the video generation, under the Frame section, select First and upload an image.

Tip

After selecting First, use the Current frame option to use the frame where the playhead is positioned as the start frame of the generated video. For best results, ensure your first frame images are visually compatible with the action you describe in your text prompt.

Select Generate.

Drag the generated video from the Generation history panel and drop it onto your timeline.

Your generated video is now part of your project and can be edited, trimmed, or combined with other media clips in the timeline.