
bytedance / seedance2-5
Seedance 2.5 is ByteDance’s next-generation AI video model for up to 30-second generation, multimodal references, and precise local editing.
Kinovi Directory
Compare published models, then open a model page for its inputs, Studio access, and API status.
52 models
Provider

Seedance 2.5 is ByteDance’s next-generation AI video model for up to 30-second generation, multimodal references, and precise local editing.

Seedance 2.5 Flash turns scene descriptions into videos, with prompts for the subject, action, camera movement, lighting, and mood. Add optional image, video, or audio references to guide the result, and create 4–30-second clips at native 480p or 720p with optional generated audio.

Next-generation AI video model by ByteDance. Combines Fast and Pro modes for high-quality video generation.

Wan 3.0 Prime multimodal reference-to-video — faster inference. Images, video, audio, web pages, and documents into 480P–1080P clips — up to 30s output without reference video; with videoUrls, reference total + output must stay within 30s.

Seedance 2.5 Flash Video Edit transforms existing footage through text-guided changes to subjects, settings, lighting, and visual style. Add optional image and audio references to guide the edit. On Kinovi, output uses native 480p or 720p, with duration and aspect ratio following the source video.

Seedance 2.5 Flash Video Extend creates new footage from the ending of an existing clip. Use a source video and a prompt to direct the next action, camera movement, and scene progression, with optional image and audio references. Add 4–30 seconds at native 480p or 720p; framing follows the source video.

Alibaba Wan 3.0 multimodal reference-to-video on Kinovi. Blend images, video, audio, web pages, and documents into 480P–1080P clips — up to 30s output without reference video; with videoUrls, reference total + output must stay within 30s.

Text-to-image, reference-guided editing, and multilingual text rendering for design-grade image work.

Advanced AI image generation and editing by OpenAI. GPT Image 2 excels at photorealism, text rendering, complex compositions, and precise instruction following.

Seedream 5.5 generates an image from a text prompt. Upstream model seedream-v5.5. Resolution: 1k / 1.5k / 2k (default 2k).

Seedance 2.0 Fast is Kinovi's lower-credit Seedance route for fast text-to-video and image-to-video iteration at 480p or 720p.

Seedream 5.5 edits or fuses up to 10 reference images with a text instruction. Console path seedream-v5.5/edit; API model is still seedream-v5.5 with image. Resolution: 1k / 1.5k / 2k (default 2k).

Seedance 2.0 Mini is a low-cost video generation model built for fast drafting, batch iteration, e-commerce assets, short-video concepts, and reference-based creative workflows.

Seedance 2.5 Video Edit restyles or locally edits an existing clip. Upload a reference video and describe the change—subject, background, lighting, or style. Output length follows the source automatically.

Seedance 2.5 Video Extends continues an existing clip. Upload a reference video, describe how the motion should continue, and choose how many seconds to add. Aspect ratio stays adaptive to the source.

Latest Midjourney model with photorealistic quality and creative control.

Create coherent cinematic videos from text with flexible 2–30 second duration, 480p–1080p output, multiple aspect ratios, and optional generated audio.

Animate a first-frame image into a coherent cinematic video with optional last-frame guidance, 2–30 second duration, 480p–1080p output, and optional generated audio.

Generate coherent cinematic videos from text with accelerated Wan 3.0 Prime inference, flexible 2–30 second duration, 480p–1080p output, multiple aspect ratios, and optional generated audio.

Animate a first-frame image with accelerated Wan 3.0 Prime inference, optional last-frame guidance, 2–30 second duration, 480p–1080p output, and optional generated audio.

HappyHorse 1.0 is Alibaba's high-quality video generation model on Kinovi, supporting text-to-video, image-to-video, reference-image video, and video editing modes.

Specialized anime and illustration AI model with exceptional Japanese art style quality.

Fast, high-quality AI image generation with creative control and photorealistic output.

SeedVR2 Image Upscale is announced on Kinovi. Studio and public REST API access are not live yet.

OpenAI GPT Image 2.5 Sunburst turns prompts into high-fidelity images. The precision tier for text-to-image and editing when detail matters most, with usage-based billing and the same controls as Flare.

Grok Imagine Video v1.5 Text to Video turns natural-language prompts into short, stylized AI videos, with 480P, 720P, and 1080P output options plus selectable aspect ratios for social media clips, creative storytelling, concept videos, marketing content, and professional text-to-video workflows. Ready-to-use REST inference API, best performance, no coldstarts, affordable pricing.

OpenAI GPT Image 2.5 Flare turns prompts into production-ready images. The speed-tuned tier for everyday text-to-image and editing, with usage-based billing and the same quality and resolution controls as Sunburst.

Grok Imagine Video v1.5 Image-to-Video turns a single input image and a natural-language prompt into short, stylized AI videos, with 480P, 720P, and 1080P output options and selectable aspect ratios for character animation, product videos, social media content, creative storytelling, and marketing workflows. Ready-to-use REST inference API, best performance, no coldstarts, affordable pricing.

Grok Imagine Video v1.5 Reference to Video turns up to seven reference images and a natural-language prompt into short, stylized AI videos, with 480P / 720P output options and selectable aspect ratios for identity-consistent clips, character-driven videos, social media content, creative storytelling, and marketing workflows. Ready-to-use REST inference API, best performance, no coldstarts, affordable pricing.

High-quality AI image generation and editing with 1K, 2K, and 4K output.

Grok Imagine Video 1.5 Lite generates videos from text descriptions using xAI's Grok Imagine Video model. Create high-quality videos with customizable duration, aspect ratio, and resolution. Ready-to-use REST inference API, best performance, no coldstarts, affordable pricing.

Prompt-only text-to-video. MP4 at 720p or 1080p; length 3–30 seconds; native audio.

Grok Imagine Video 1.5 Lite transforms images into videos using xAI's Grok Imagine Video model. Animate still images with natural motion, scene continuity, and synchronized audio. Ready-to-use REST inference API, best performance, no coldstarts, affordable pricing.

Keyframe-driven image-to-video. Animate a first frame, or move through up to 10 keyframes in order. Output keeps the input aspect ratio. MP4 at 720p or 1080p; 3–30 seconds; native audio.

AI music generation with Suno.

Reference-driven video. Guide the clip with reference images, reference videos or keyframes, or any mix of them. MP4 at 720p or 1080p; 3–30 seconds; native audio.

Image-to-video from a first-frame image (optional last frame). MP4 at 480p or 768p; length 5–15 seconds. Output aspect ratio follows the input image.

Prompt-only text-to-video. MP4 at 480p or 768p; length 5–15 seconds.
Talking-head video from one portrait and one audio track. Lip-synced 480p or 720p output, 2–15 seconds.

AI sample generation with Suno.

AI remix with Suno.

AI sound effects with Suno.

Prompt-only text-to-video. MP4 at 768p or 2k; length 4–15 seconds.

Image-to-video from a first-frame image (optional last frame). MP4 at 768p or 2k; length 4–15 seconds. Output aspect ratio follows the input image.

Reference-guided video from images, videos, and/or audio (at least one). MP4 at 768p or 2k; length 4–15 seconds. Not combinable with first/last frames.

Prompt-only text-to-video. MP4 at 480p or 768p; length 5–15 seconds.

Image-to-video from a first-frame image (optional last frame). MP4 at 480p or 768p; length 5–15 seconds. Output aspect ratio follows the input image.

Reference-guided video from images, videos, and/or audio (at least one). MP4 at 480p or 768p; length 5–15 seconds. Not combinable with first/last frames.

Use seed-audio-1-0-voice-text-to-voice on Kinovi to run ByteDance Seed Audio 1.0 text to voice from a single prompt. Describe timbre, emotion, dialogue, and scene; add optional ambience and SFX. Each pass can reach about two minutes across 20+ languages, with roughly 100 ms dialogue timing cues. Call the live model by Model API ID seed-audio-1-0-voice-text-to-voice via the Kinovi API or Studio.

ByteDance's Seed Audio 1.0 reference-audio entry on the Kinovi API. Upload up to 3 clean speech clips and cite them as @Audio1–@Audio3 to clone timbre for new lines or continuations, with scene-level performance. Cannot combine with an image reference.

ByteDance's Seed Audio 1.0 reference-image entry on the Kinovi API. Upload one person or character image; the model infers a matching voice and speaks your prompt. Image and audio references are exclusive.

ByteDance's Seed Audio 1.0 sound-design entry on the Kinovi API. Text-to-SFX and ambience beds — rain alleys, cafés, Foley, whoosh/impact stingers — oriented to non-dialogue-first audio from the same full-scene model.