Video Models Supporting First and Last Frame

Runware supports a range of video models that can use separate first and last frame images to guide a generation.

Some currently supported options include:

ByteDance

• Seedance 1.0 Pro

• Seedance 1 Lite

• Seedance 1.5 Pro

• Seedance 2.0

Seedance 2.0 supports up to two frameImages, which are automatically treated as the first and last frames when two images are provided.

Google Veo

• Veo 2

• Veo 3

• Veo 3 Fast

• Veo 3.1

• Veo 3.1 Fast

• Veo 3.1 Lite

Veo 3.1 and its variants support first and last frame inputs through frameImages.

Kling AI

• Kling 2.1 Pro

• Kling 2.5 Turbo Pro

• Kling 2.6 Pro

• Kling 3.0 Standard

• Kling 3.0 Pro

• Kling O1 Standard

• Kling O1 Pro

• Kling O3 Standard

• Kling O3 Pro

Kling models that support this workflow accept up to two frame images for first and last frame guidance.

Vidu

• Vidu Q1

• Vidu Q1 Classic

• Vidu Q2 Turbo

• Vidu Q2 Pro

• Vidu 1.5

• Vidu 2.0

• Vidu Q3

• Vidu Q3 Turbo

Supported Vidu models accept two frame images, with the first image used as the opening frame and the second as the final frame.

PixVerse

• PixVerse v3.5

• PixVerse v4

• PixVerse v4.5

• PixVerse v5

• PixVerse v5.5

• PixVerse v5.6

PixVerse v5.6, for example, supports explicit first and last frame positions.

Wan

• Wan 3.0

Wan 3.0 supports up to two frameImages, including first and last frame generation.

Find the latest supported models

As new video models are added regularly, the easiest way to find the latest options is to use the Runware Models catalog and select:

Video → An image → As the first and last frame

https://runware.ai/docs/models

You can also open an individual model page to confirm its supported inputs, parameters, resolutions, durations, and other model-specific requirements.