Private AI

Nano GPT logo
NanoGPT
Nano GPT logo
NanoGPT
Nano GPT logo
NanoGPT
Search
 
 
 
 
All conversations

Create

Chat
Studio

Explore

Models
Pricing
AI Apps
Blog

Library

Gallery
Workspace
Usage & billing
API
Suggestions and bugs
Use cases

|

Terms

|

Privacy

|

Refunds
Usage & billing
Teams
Referrals
Presets
Settings
FAQPress & mediaBrand assetsBenchmarksNano RPCEarn crypto
  1. All models overview
  2. Video
  3. Skywork Ai

SkyReels V4

Try Model

Skywork AI SkyReels V4 unified video model for text-to-video, image-to-video, and reference-to-video generation with optional synchronized sound. Supports 480p, 720p, or 1080p output and 3-15 second clips.

Added May 24, 2026

Details

Starting Price
From $0.240 per video
Model type
both

Final price depends on the selected settings and is shown before generation.

Settings

Generation controls available for this model.

Output Format

up to 1,080p
Square
Portrait
Landscape

Default Duration

5

13 duration options

Aspect Ratio

Select

Default

16:9

Options (3)

16:9, 9:16, 1:1

Output aspect ratio for text and reference modes

Duration

Select

Default

5

Options (13)

3 seconds, 4 seconds, 5 seconds, 6 seconds +9 more

3-15 seconds

Generate Sound

Toggle

Default

No

Enable synchronized sound output. Fast mode disables sound.

Quality Mode

Select

Default

std

Options (2)

Standard, Fast

Standard or faster generation

Resolution

Select

Default

1080p

Options (3)

480p, 720p, 1080p

Output resolution

Benchmarks

Human preference benchmarks sourced from LMArena.

Text to Video

#9 / 50

Arena Score

1458.7

Votes

1,060

Confidence Interval

1440.4 - 1477.0

Image to Video

#7 / 50

Arena Score

1475.3

Votes

1,542

Confidence Interval

1459.7 - 1491.0

Video Edit

#7 / 12

Arena Score

1363.5

Votes

697

Confidence Interval

1342.5 - 1384.5

Published 2026-11-10 · Matched as SkyReels V4 · CC BY 4.0

LMArena Dataset

Examples

Loading examples…

Related video models

Compare SkyReels V4 with similar models from the same provider or model family.

MiniMax H3 Hyper

minimax/h3-hyper

MiniMax H3 Hyper is an eight-evaluation distilled H3 variant for faster video creation with native stereo audio. Generate from text or animate a starting image with an optional end frame at 768p and 24 fps. Request 4–15 whole seconds; clips round up to the model’s frame grid (about 4.46–15.08 seconds). Reference mode supports six-second clips with one or three images. Text and reference modes offer six aspect ratios; image animation follows the starting image. Detailed shot and sound descriptions work best. Video and audio references, seeds, and prompt expansion are not supported.

HeyGen Avatar 6

heygen/avatar-6

Create a talking video from speech audio and one of 1,286 stock avatars, a portrait image, or an animated-character reference video. Choose 1080p or 2K output, with optional motion directions. Stock avatars support five aspect ratios; portrait and reference modes preserve source framing. Upload WAV or MP3 speech up to 8 seconds and 25 MiB, or JPEG, PNG or WebP portraits up to 20 MiB. Reference videos must show animated or illustrated characters in MP4, up to 10 seconds and 25 MiB; real-person reference videos are unsupported. Script-to-speech is unavailable. Final billing rounds generated-video duration up to whole seconds, with a one-second minimum. Reserved up front, unused refunded. Pricing is $0.10 per generated second at 1080p or $0.145 at 2K.

HeyGen Avatar 6 Turbo

heygen/avatar-6/turbo

Create a talking video from speech audio and one of 1,286 stock avatars, a portrait image, or an animated-character reference video. Choose 1080p or 2K output, with optional motion directions. Stock avatars support five aspect ratios; portrait and reference modes preserve source framing. Upload WAV or MP3 speech up to 8 seconds and 25 MiB, or JPEG, PNG or WebP portraits up to 20 MiB. Reference videos must show animated or illustrated characters in MP4, up to 10 seconds and 25 MiB; real-person reference videos are unsupported. Script-to-speech is unavailable. Final billing rounds generated-video duration up to whole seconds, with a one-second minimum. Reserved up front, unused refunded. Turbo costs $0.05 per generated second at 1080p or $0.0875 at 2K.

LongCat Avatar 1.6

wavespeed-ai/longcat-avatar-1.6

Animate one person or two people from an image and speech or singing audio. Supports lip-synced talking and singing, optional expression prompts, simultaneous or sequential two-person dialogue, and 480p or 720p output. Clips are limited to 64 seconds; longer audio is trimmed.

Seedance 2.5 Flash

bytedance/seedance-2.5-flash

Seedance 2.5 Flash creates 4–30 second videos from text, start and end images, or visual references, and edits existing clips at 480p or 720p with synchronized audio. Video extension appends 4–30 new seconds to a source clip and returns the joined video. Motion control transfers movement from a 4–30 second source video to a character described in the prompt or reference images.

3D to Video

wavespeed-ai/3d-to-video

Turn a rough 3D viewport animation or blocking playblast into a finished shot. Supply a layout video, a restyled first-frame image and a prompt describing the finished look; optionally add a restyled last frame and a seed. The layout guides camera movement, motion and object placement, while the images guide appearance. Choose 720p or 1080p, preserving the input aspect ratio. Processes the first 120 seconds at up to 30 FPS; timing and dimensions may vary. Does not generate audio. Billing uses input duration rounded up to whole seconds, with a 3-second minimum and 120-second maximum.