Compare workflows. Watch the results.

Compare video, avatar and image models available on Seedance3.app. The specifications below describe our current tools; check the generator for settings, access and Star costs.

AI video

AI video

Seedance 2.5

Multimodal video workflows

Use Seedance 2.5 for 4–30 second clips with text, first and last frames, or multimodal references. Flash supports 480p and 720p across all account tiers. Choose High for 1080p output; High requires paid membership on Seedance3.app.

Reference inputs
Text prompt / Image / Video / Audio
Best for
Prompt to video / First and last frames / Multimodal references
Available tiers
ModelResolution by tierDuration
Flash480p / 720p4–30 Seconds
HighPaid membership480p / 720p / 1080p4–30 Seconds
View original ↗

AI video

Seedance 2.0

Everyday video creation

Use Seedance 2.0 for 4–15 second scenes from text, first and last frames, or image, video and audio references. Compare Fast and Mini when 480p or 720p fits your project; choose High when you need the 1080p or 4K options available here.

Reference inputs
Text prompt / Image / Video / Audio
Best for
Prompt to video / First and last frames / Multimodal references
Available tiers
ModelResolution by tierDuration
Fast480p / 720p4–15 Seconds
Mini480p / 720p4–15 Seconds
High480p / 720p / 1080p / 4K4–15 Seconds
View original ↗

AI video

HappyHorse 1.1

Character and reference-led scenes

Use HappyHorse 1.1 for 3–15 second character scenes, product presentations or short stories guided by images and reference video. Choose 720p or 1080p for your delivery needs. This workflow does not generate audio, so plan voice-over or music in your editing stage.

Reference inputs
Text prompt / Image / Video
Best for
Prompt to video / First and last frames / Multimodal references
Available tiers
ModelResolution by tierDuration
1.1720p / 1080p3–15 Seconds
View original ↗

AI video

Wan 3.0 Prime

Longer reference-led sequences

Consider Wan 3.0 Prime when a reference-led scene needs up to 30 seconds. It accepts image, video and audio references and offers 480p, 720p and 1080p here. Prime requires paid membership; compare Seedance 2.5 Flash if you need a 30-second option open to all account tiers.

Reference inputs
Text prompt / Image / Video / Audio
Best for
Prompt to video / First and last frames / Multimodal references
Available tiers
ModelResolution by tierDuration
PrimePaid membership480p / 720p / 1080p2–30 Seconds

AI video

Kling 4.0

Reference-driven motion

Use Kling 4.0 with image and video references to guide product scenes or character motion. Choose Flash for 720p clips of 3–20 seconds; choose Standard when you need 1080p or a duration above 20 seconds, up to 30 seconds.

Reference inputs
Text prompt / Image / Video
Best for
Prompt to video / Multimodal references
Available tiers
ModelResolution by tierDuration
Flash720p3–20 Seconds
Standard720p / 1080p3–30 Seconds
View original ↗

AI video

Veo 3.1

First and last frame workflows

Use Veo 3.1 for short scenes from a prompt, first and last frames, or image references. Text and frame workflows offer 4, 6 or 8 seconds; image-reference mode uses 8 seconds. Lite, Fast and Quality offer 720p, 1080p and 4K here; compare the current Star quote before choosing a tier.

Reference inputs
Text prompt / Image
Best for
Prompt to video / First and last frames / Image references
Available tiers
ModelResolution by tierDuration
Fast720p / 1080p / 4K4 / 6 / 8 Seconds
Lite720p / 1080p / 4K4 / 6 / 8 Seconds
Quality720p / 1080p / 4K4 / 6 / 8 Seconds
View original ↗

Talking avatars

Talking avatars

Omnihuman 2.0

Portrait and audio to talking video

Choose Omnihuman 2.0 when you already have a portrait and a speech recording for a product explanation or presentation. It turns the portrait and audio into a lip-synced 720p video. Prepare your script and audio first, then use this portrait-and-audio workflow for the talking segment.

Reference inputs
Image / Audio
Best for
Multimodal references
Available tiers
ModelResolution by tierDuration
2.0720p4–300 Seconds
View original ↗

AI images

AI images

GPT Image

Image editing and visual concepts

Use GPT Image to create product visuals, concept images or edits from a prompt and reference images. Choose GPT Image 2 or a 2.5/2.6 variant with the output size you need. The 2.6 eco option is limited to 1K here; select a variant listing 2K or 4K when you need those sizes.

Reference inputs
Text prompt / Image
Available tiers
ModelResolution by tier
gpt image 21K / 2K / 4K
gpt image 2.6 eco1K
gpt image 2.6 sunburst1K / 2K / 4K
gpt image 2.6 flare1K / 2K / 4K
gpt image 2.5 flare1K / 2K / 4K
gpt image 2.5 sunburst1K / 2K / 4K

AI images

Gemini & Nano Banana

Image generation and visual exploration

Use Gemini and Nano Banana for prompt-led image concepts and variations guided by reference images. Nano Banana 2.1, Gemini 3.1 and Gemini 3 Pro offer 1K, 2K and 4K here. Gemini 3.1 Lite is limited to 1K; choose another listed variant when you need a larger output.

Reference inputs
Text prompt / Image
Available tiers
ModelResolution by tier
Nano Banana 2.11K / 2K / 4K
gemini 3.11K / 2K / 4K
gemini 3 pro1K / 2K / 4K
gemini 3.1 lite1K

AI images

Seedream

Prompt-led images and variations

Use Seedream 5.0 Flash or 5.5 for product images, visual concepts and variations based on text or reference images. Both offer 1K and 2K here, so compare their current Star quotes and your project results when selecting a version. Choose a model listing 4K if that is your required output size.

Reference inputs
Text prompt / Image
Available tiers
ModelResolution by tier
seedream 5.51K / 2K
seedream 5.0 flash1K / 2K

AI images

MiniMax H3 Image

Fast image ideas and iterations

Use MiniMax H3 Image for still product visuals, concept images and reference-led variations. It offers 1K, 2K and 4K here; choose the output size and aspect ratio for your publishing format. For an animated product shot, select a video workflow after preparing your image.

Reference inputs
Text prompt / Image
Available tiers
ModelResolution by tier
minimax h31K / 2K / 4K