Model catalog
Video models
Browse every available video model, organized by family and ready to open with the right creation workflow.
162 models33 families
LTX Video
Lightricks · 23 models
LightricksLTX 2.3LTX 2.3 text-to-video — high-resolution cinematic clips up to 4K with native audio generation and resolution-scaled pricing.LightricksNewLTX 2.5 Fast (Audio to Video)Fast LTX 2.5 audio-driven video generation from a 2–20 second audio clip, with an optional first frame and optional visual prompt.LightricksNewLTX 2.5 Fast (Image to Video)Fast LTX 2.5 image-to-video with optional end-frame transitions, 6–20 second output up to 4K, and native audio.LightricksNewLTX 2.5 Fast (Text to Video)Fast LTX 2.5 text-to-video with 6–20 second output up to 4K, synchronized native audio, and camera-motion control.LightricksNewLTX 2.5 Pro (Audio to Video)LTX 2.5 Pro audio-driven video generation from a 2–10 second audio clip, with an optional first frame and optional visual prompt.LightricksNewLTX 2.5 Pro (Image to Video)LTX 2.5 Pro image-to-video with optional end-frame transitions, synchronized native audio, and camera-motion control.LightricksNewLTX 2.5 Pro (Text to Video)LTX 2.5 Pro text-to-video with cinematic 720p/1080p output, synchronized native audio, and camera-motion control.LightricksNewLTX-2 19B (I2V)LTX-2 19B image-to-video — animate a start image into ultra-high-resolution clips up to 4K with optional end-frame interpolation and native audio.LightricksNewLTX-2 19B (T2V)LTX-2 19B text-to-video — ultra-high-resolution generation up to 4K with camera LoRA controls and native audio synthesis.LightricksNewLTX-2.3 (I2V LoRA)LTX-2.3 22B image-to-video with LoRA — animate a start image using a custom fine-tune for stylized motion, with optional end-frame support.LightricksNewLTX-2.3 (T2V LoRA)LTX-2.3 22B text-to-video with LoRA — apply custom fine-tune weights from Hugging Face to steer the visual style of generated videos.LightricksNewLTX-2.3 (V2V LoRA)LTX-2.3 22B reference-video-to-video with LoRA — restyle a reference video with a custom fine-tune using depth, canny, or pose preprocessing.LightricksNewLTX-2.3 (V2V Reference)LTX-2.3 22B reference-video-to-video — restyle or transform an existing video using depth, canny, or pose preprocessing with native audio generation.LightricksNewLTX-2.3 22B (I2V)LTX-2.3 22B image-to-video — animate a start image with optional end-frame using the full 22B model, with native audio and advanced scheduler controls.LightricksNewLTX-2.3 22B (T2V)LTX-2.3 22B text-to-video — flagship Lightricks model with multi-scale generation, advanced scheduler options, and native audio synthesis.LightricksNewLTX-2.3 Distilled (I2V LoRA)Distilled LTX-2.3 22B image-to-video with LoRA support — animate a start image using a custom fine-tune with fast distilled generation and audio synthesis.LightricksNewLTX-2.3 Distilled (I2V)Distilled LTX-2.3 22B image-to-video — animate a start image into a smooth clip up to 1440p with camera LoRA control and native audio.LightricksNewLTX-2.3 Distilled (T2V LoRA)Distilled LTX-2.3 22B text-to-video with LoRA support — combine a custom fine-tune with fast distilled generation and native audio synthesis.LightricksNewLTX-2.3 Distilled (T2V)Distilled LTX-2.3 22B text-to-video — fast, high-fidelity generation up to 1440p with native audio synthesis.LightricksNewLTX-2.3 Distilled (V2V LoRA)Distilled LTX-2.3 22B video-to-video with LoRA support — apply a custom fine-tune style while restyling reference footage with audio generation.LightricksNewLTX-2.3 Distilled (V2V)Distilled LTX-2.3 22B video-to-video — restyle reference footage faster and cheaper than full V2V, with depth/canny/pose preprocessors and audio output.LightricksNewLTX-2.3 I2V FastLTX-2.3 image-to-video fast endpoint — animate a start image into a 6–20 second clip up to 4K with optional end frame and native audio.LightricksNewLTX-2.3 T2V FastLTX-2.3 text-to-video fast endpoint — ultra-quick generation up to 4K with 6–20 second clips and native audio synthesis.
WAN
Alibaba · 17 models
AlibabaNewWan 2.2 (I2V LoRA)WAN 2.2 A14B image-to-video with LoRA — animate a start image with a custom fine-tune for stylized motion, including optional end-frame control.AlibabaNewWan 2.2 (T2V LoRA)WAN 2.2 A14B text-to-video with LoRA — apply custom fine-tune weights to steer the visual style of generated videos.AlibabaNewWAN 2.2 A14B Turbo (i2v)WAN 2.2 A14B Turbo image-to-video — flat-rate accelerated i2v with optional end-frame support across three resolution tiers.AlibabaNewWAN 2.2 Animate (Mix)WAN 2.2 Animate Mix — replace the performer in a source video with a character from a reference image, keeping the original motion, scene, and camera.AlibabaNewWAN 2.2 Animate (Move)WAN 2.2 Animate Move — transfer the motion of a source video onto a static character image, animating the character with the driving performance.AlibabaNewWan 2.2 Turbo (T2V)WAN 2.2 A14B Turbo text-to-video — accelerated generation with quality and write-mode controls for fast iteration at lower cost.AlibabaNewWAN 2.7 (Edit)WAN 2.7 video editing — transform an input video by style, effect, or camera replication while preserving the original motion, with full resolution and duration controls.AlibabaNewWAN 2.7 (I2V)WAN 2.7 image-to-video — animate a start frame, or first-and-last frames, or continue from a video, into a smooth cinematic clip with full resolution and duration controls.AlibabaNewWAN 2.7 (Reference)WAN 2.7 reference-to-video — place reference characters into a prompted scene, or drive a video-to-video transformation, with identity consistency and full resolution/duration controls.AlibabaNewWAN 2.7 (T2V)WAN 2.7 text-to-video — Alibaba's latest generation video model with aspect ratio, resolution, and duration controls plus optional custom audio.
AlibabaNewWAN 3.0 (Image to Video)WAN 3.0 image-to-video using a first frame and optional last frame, with native audio and up to 30 seconds of motion.
AlibabaNewWAN 3.0 (Reference to Video)WAN 3.0 multimodal reference generation using images, video, and audio for consistent characters, objects, motion, and sound.
AlibabaNewWAN 3.0 (Text to Video)WAN 3.0 text-to-video with native audio, adaptive duration, cinematic motion, and clips up to 30 seconds.
AlibabaNewWAN 3.0 (Video Edit)WAN 3.0 video editing and replication with optional image and audio references while preserving temporal consistency.AlibabaNewWAN 3.0 Prime (Image)WAN 3.0 Prime animates a required start frame and optional end frame into a 2–30 second video with native audio.AlibabaNewWAN 3.0 Prime (Reference)WAN 3.0 Prime combines up to ten images, five videos, and five audio references into a controlled 2–30 second video.AlibabaNewWAN 3.0 Prime (Text)WAN 3.0 Prime creates 2–30 second videos from text with native audio, prompt expansion, and output up to 1080p.
AlibabaNewWAN 3.0 (Image to Video)WAN 3.0 image-to-video using a first frame and optional last frame, with native audio and up to 30 seconds of motion.AlibabaNewWAN 3.0 (Reference to Video)WAN 3.0 multimodal reference generation using images, video, and audio for consistent characters, objects, motion, and sound.AlibabaNewWAN 3.0 (Text to Video)WAN 3.0 text-to-video with native audio, adaptive duration, cinematic motion, and clips up to 30 seconds.AlibabaNewWAN 3.0 (Video Edit)WAN 3.0 video editing and replication with optional image and audio references while preserving temporal consistency.AlibabaNewWAN 3.0 Prime (Image)WAN 3.0 Prime animates a required start frame and optional end frame into a 2–30 second video with native audio.AlibabaNewWAN 3.0 Prime (Reference)WAN 3.0 Prime combines up to ten images, five videos, and five audio references into a controlled 2–30 second video.AlibabaNewWAN 3.0 Prime (Text)WAN 3.0 Prime creates 2–30 second videos from text with native audio, prompt expansion, and output up to 1080p.Flux
Black Forest Labs · 13 models
Black Forest LabsNewFLUX Video UpscaleBlack Forest Labs FLUX Video Upscale increases source-video resolution while preserving motion and optionally reconstructing detail from a prompt.
Black Forest LabsNewFLUX.3 Draft (I2V)Animate an opening image as a fast FLUX.3 draft and retain the cache for a matching full-quality render.
Black Forest LabsNewFLUX.3 Draft (T2V)Generate a fast, low-cost FLUX.3 text-to-video preview and retain its encrypted cache for a matching full-quality render.
Black Forest LabsNewFLUX.3 Draft EnhanceRender an approved FLUX.3 encrypted draft cache at full quality while preserving the draft's subjects, composition, and motion. ArtEmotion derives billing from owned drafts; external caches are billed at the 20-second maximum.
Black Forest LabsNewFLUX.3 Draft ExtendQuickly preview a continuation of a source MP4 and retain its cache for a matching full-quality enhancement.
Black Forest LabsNewFLUX.3 Draft First + LastPreview the exact motion between required opening and closing images, then enhance the approved draft without changing its composition.
Black Forest LabsNewFLUX.3 Draft KeyframesPreview a controlled FLUX.3 shot from up to ten ordered keyframes, with optional exact positions at 24 fps.
Black Forest LabsNewFLUX.3 Edit VideoEdit a source video with natural-language instructions while preserving its motion, timing, and framing. MP4 input under 15 seconds and 50 MB; billed by the source clip's measured duration.
Black Forest LabsNewFLUX.3 Extend VideoContinue a source MP4 beyond its final frame while preserving motion, framing, dialogue, and audio continuity.
Black Forest LabsNewFLUX.3 First + Last FrameCreate a FLUX.3 video that moves coherently between required opening and closing images, with exact duration control and native audio.
Black Forest LabsNewFLUX.3 KeyframesControl a FLUX.3 shot with up to ten ordered keyframe images and optional exact frame positions at 24 fps.
Black Forest LabsNewFLUX.3 Video (I2V)Animate an opening image into a controlled FLUX.3 video with native synchronized audio and up to 20 seconds of motion.
Black Forest LabsNewFLUX.3 Video (T2V)Black Forest Labs FLUX.3 text-to-video with native audio, multilingual dialogue, multi-shot prompting, and clips up to 20 seconds.
Black Forest LabsNewFLUX.3 Draft (I2V)Animate an opening image as a fast FLUX.3 draft and retain the cache for a matching full-quality render.Black Forest LabsNewFLUX.3 Draft (T2V)Generate a fast, low-cost FLUX.3 text-to-video preview and retain its encrypted cache for a matching full-quality render.Black Forest LabsNewFLUX.3 Draft EnhanceRender an approved FLUX.3 encrypted draft cache at full quality while preserving the draft's subjects, composition, and motion. ArtEmotion derives billing from owned drafts; external caches are billed at the 20-second maximum.Black Forest LabsNewFLUX.3 Draft ExtendQuickly preview a continuation of a source MP4 and retain its cache for a matching full-quality enhancement.Black Forest LabsNewFLUX.3 Draft First + LastPreview the exact motion between required opening and closing images, then enhance the approved draft without changing its composition.Black Forest LabsNewFLUX.3 Draft KeyframesPreview a controlled FLUX.3 shot from up to ten ordered keyframes, with optional exact positions at 24 fps.Black Forest LabsNewFLUX.3 Edit VideoEdit a source video with natural-language instructions while preserving its motion, timing, and framing. MP4 input under 15 seconds and 50 MB; billed by the source clip's measured duration.Black Forest LabsNewFLUX.3 Extend VideoContinue a source MP4 beyond its final frame while preserving motion, framing, dialogue, and audio continuity.Black Forest LabsNewFLUX.3 First + Last FrameCreate a FLUX.3 video that moves coherently between required opening and closing images, with exact duration control and native audio.Black Forest LabsNewFLUX.3 KeyframesControl a FLUX.3 shot with up to ten ordered keyframe images and optional exact frame positions at 24 fps.Black Forest LabsNewFLUX.3 Video (I2V)Animate an opening image into a controlled FLUX.3 video with native synchronized audio and up to 20 seconds of motion.Black Forest LabsNewFLUX.3 Video (T2V)Black Forest Labs FLUX.3 text-to-video with native audio, multilingual dialogue, multi-shot prompting, and clips up to 20 seconds.Kling
Kuaishou · 13 models
KuaishouNewKling 2.5 Turbo (I2V)Kling 2.5 Turbo image-to-video — fast 1080p animation from a start image with optional end-frame.KuaishouNewKling 2.5 Turbo (T2V)Kling 2.5 Turbo text-to-video — max creativity with exceptional value. Fast 1080p generation.KuaishouNewKling 2.6 (I2V)Kling 2.6 image-to-video — animate a start image with strong audio-visual synchronisation and optional end frame.KuaishouNewKling 2.6 (T2V)Kling 2.6 text-to-video — see the sound, hear the visual. Specialised in audio-visual synchronisation.KuaishouNewKling 2.6 Motion ControlKling 2.6 Motion Control — apply human motion from a reference video onto a character in a reference image. Output duration matches the reference video (3–30s).KuaishouNewKling 3.0 (I2V)Kling 3.0 image-to-video — animate a start image up to 4K with optional end-frame and native audio.KuaishouNewKling 3.0 (T2V)Kling 3.0 text-to-video — native audio, multi-shot storytelling, and strong element consistency up to 4K.KuaishouNewKling 3.0 Motion ControlKling 3.0 Motion Control — apply human motion from a reference video onto a character in a reference image. Higher fidelity than 2.6. Output duration matches the reference video (3–30s).KuaishouNewKling 3.0 Omni (I2V)Kling 3.0 Omni image-to-video — multimodal input animates a reference image with start/end frame control, reference video, and native audio.KuaishouNewKling 3.0 Omni (T2V)Kling 3.0 Omni text-to-video — multimodal generation with reference images, reference video, native audio, and multi-shot storyboards.KuaishouNewKling 3.0 Turbo (I2V)Kling 3.0 Turbo image-to-video — animate a start image via the new Turbo API (first frame only).KuaishouNewKling 3.0 Turbo (T2V)Kling 3.0 Turbo text-to-video — consistently excellent results at a better cost point via the new Turbo API.KuaishouNewKling AvatarKling Avatar — bring a portrait to life with synchronized speech and expressive motion driven by an audio file.
MiniMax
MiniMax · 13 models
MiniMaxNewMiniMax H3 Max (Image)MiniMax H3 Max animates an optional first and last frame into a 5–15 second video at 480P or 768P.MiniMaxNewMiniMax H3 Max (Reference)MiniMax H3 Max reference-to-video combines image, video, and audio references into a consistent 5–15 second scene.MiniMaxNewMiniMax H3 Max (Text)MiniMax H3 Max text-to-video with 5–15 second clips, cinematic prompt expansion, and 480P or 768P output.
MiniMaxNewMiniMax H3 Max Camera ControlsAnimate a still image with keyframed camera orbit, elevation, and distance. Preserve the scene while directing camera motion, including signed full turns and 1080P refinement.
MiniMaxNewMiniMax H3 Max LoRA (Image)MiniMax H3 Max image-to-video with optional first and last frames plus one to three custom LoRA adapters at 480P or 768P.
MiniMaxNewMiniMax H3 Max LoRA (Text)MiniMax H3 Max text-to-video with one to three custom LoRA adapters, cinematic prompt expansion, and 5–15 second output at 480P or 768P.
MiniMaxNewMiniMax H3 Max Multi AngleAnimate a still image with keyframed camera orbit, elevation, and distance. Preserve the scene while directing camera motion, including signed full turns and 1080P refinement.
MiniMaxNewMiniMax H3 Max Turbo (Image)MiniMax H3 Max Turbo animates an optional first and last frame with improved prompt adherence and faster 480P or 768P generation.
MiniMaxNewMiniMax H3 Max Turbo (Text)MiniMax H3 Max Turbo text-to-video with stronger prompt adherence, high-throughput generation, and 5–15 second output at 480P or 768P.MiniMaxNewMiniMax Hailuo-02 (I2V)MiniMax Hailuo-02 Standard image-to-video — animate a start image to a 6 or 10 second cinematic clip.MiniMaxNewMiniMax Hailuo-02 (T2V)MiniMax Hailuo-02 Standard text-to-video — 6 or 10 second cinematic clips with prompt optimizer.MiniMaxNewMiniMax Hailuo-02 Pro (I2V)MiniMax Hailuo-02 Pro image-to-video — high-quality 1080p image animation with optional end frame.MiniMaxNewMiniMax Hailuo-02 Pro (T2V)MiniMax Hailuo-02 Pro text-to-video — high-quality 1080p generation with prompt optimizer.
MiniMaxNewMiniMax H3 Max Camera ControlsAnimate a still image with keyframed camera orbit, elevation, and distance. Preserve the scene while directing camera motion, including signed full turns and 1080P refinement.MiniMaxNewMiniMax H3 Max LoRA (Image)MiniMax H3 Max image-to-video with optional first and last frames plus one to three custom LoRA adapters at 480P or 768P.MiniMaxNewMiniMax H3 Max LoRA (Text)MiniMax H3 Max text-to-video with one to three custom LoRA adapters, cinematic prompt expansion, and 5–15 second output at 480P or 768P.MiniMaxNewMiniMax H3 Max Multi AngleAnimate a still image with keyframed camera orbit, elevation, and distance. Preserve the scene while directing camera motion, including signed full turns and 1080P refinement.MiniMaxNewMiniMax H3 Max Turbo (Image)MiniMax H3 Max Turbo animates an optional first and last frame with improved prompt adherence and faster 480P or 768P generation.MiniMaxNewMiniMax H3 Max Turbo (Text)MiniMax H3 Max Turbo text-to-video with stronger prompt adherence, high-throughput generation, and 5–15 second output at 480P or 768P.MiniMaxNewMiniMax Hailuo-02 (I2V)MiniMax Hailuo-02 Standard image-to-video — animate a start image to a 6 or 10 second cinematic clip.MiniMaxNewMiniMax Hailuo-02 (T2V)MiniMax Hailuo-02 Standard text-to-video — 6 or 10 second cinematic clips with prompt optimizer.MiniMaxNewMiniMax Hailuo-02 Pro (I2V)MiniMax Hailuo-02 Pro image-to-video — high-quality 1080p image animation with optional end frame.MiniMaxNewMiniMax Hailuo-02 Pro (T2V)MiniMax Hailuo-02 Pro text-to-video — high-quality 1080p generation with prompt optimizer.PixVerse
PixVerse · 13 models
PixVerseNewPixVerse LipsyncPixVerse Lipsync — sync a portrait video to provided audio or TTS-generated speech with natural head movement and cinematic lip sync.PixVerseNewPixVerse Sound EffectsPixVerse Sound Effects — generate and bake matching sound effects into a silent video, with the option to preserve any existing audio.PixVerseNewPixVerse SwapPixVerse Swap — replace a person, object, or background in any video by uploading a reference image, with optional original audio preservation.PixVerseNewPixVerse v4 (I2V)PixVerse v4 image-to-video — animate a start image with camera movement controls and style presets at an affordable rate.PixVerseNewPixVerse v4 (T2V)PixVerse v4 text-to-video — fast, affordable video generation with style presets and 5 or 8 second clip options.PixVerseNewPixVerse v4.5 (I2V)PixVerse v4.5 image-to-video — animate a start image with camera movement controls and style presets.PixVerseNewPixVerse v4.5 (T2V)PixVerse v4.5 text-to-video — improved PixVerse generation with style presets at a moderate price point.PixVerseNewPixVerse v5.5PixVerse v5.5 text-to-video — flat per-video pricing with style presets, thinking mode, and multi-clip support for 5, 8, or 10 second clips.PixVerseNewPixVerse v5.5 (I2V)PixVerse v5.5 image-to-video — animate a start image with style presets, thinking mode, multi-clip generation, and native audio.PixVerseNewPixVerse v6 (I2V)PixVerse v6 image-to-video — animate a start image with style presets, thinking mode, multi-clip generation, and native audio.PixVerseNewPixVerse v6 (T2V)PixVerse v6 text-to-video — cinematic generation with per-second pricing, style presets, thinking mode, and native audio synthesis.PixVerseNewPixVerse v6 ExtendPixVerse v6 Extend — continue an existing video clip forward with a prompt, maintaining visual style and optionally generating native audio.PixVerseNewPixVerse v6 TransitionPixVerse v6 Transition — generate a cinematic video that smoothly transitions from a start image to an end image, with optional audio and multi-clip support.
Seedance
ByteDance · 13 models
ByteDanceNewDreamina Seedance 2.5 (Multimodal)ByteDance's public multimodal video model for text-to-video, first/last-frame animation, multi-asset reference generation, video editing, extension, and native audio. It creates coherent clips up to 30 seconds at a fixed 24fps.ByteDanceNewSeedance 2.0 (I2V)ByteDance Seedance 2.0 image-to-video — animate a start image into a cinematic clip with synchronized native audio and optional end-frame control.ByteDanceNewSeedance 2.0 (Reference)ByteDance Seedance 2.0 reference-to-video — maintain character identity from a portrait photo across any scene, with native audio generation.ByteDanceNewSeedance 2.0 (T2V)ByteDance Seedance 2.0 text-to-video — premium cinematic video generation with native audio synthesis and support for a wide range of durations and resolutions.ByteDanceNewSeedance 2.0 (V2V)ByteDance Seedance 2.0 video-to-video — edit or extend an existing video clip with a text prompt. Repaint style, continue the scene, or remix motion at full 1080p quality.ByteDanceNewSeedance 2.0 Fast (I2V)ByteDance Seedance 2.0 Fast image-to-video — quickly animate a start image with optional end-frame guidance and native audio at reduced cost.ByteDanceNewSeedance 2.0 Fast (Reference)ByteDance Seedance 2.0 Fast reference-to-video — maintain character identity from a portrait photo across any scene at lower cost and faster speed.ByteDanceNewSeedance 2.0 Fast (T2V)ByteDance Seedance 2.0 Fast text-to-video — a faster, lower-cost tier of Seedance 2.0 for rapid prototyping with native audio and wide aspect ratio support.ByteDanceNewSeedance 2.0 Fast (V2V)ByteDance Seedance 2.0 Fast video-to-video — edit or extend an existing video clip at lower cost and faster speed. Great for rapid iteration on style transfer and scene extension.ByteDanceNewSeedance 2.0 Mini (I2V)ByteDance Seedance 2.0 Mini image-to-video — animate a start image into a short cinematic clip with native audio at the lowest cost in the Seedance 2.0 family.
ByteDanceNewSeedance 2.0 Mini (Reference)ByteDance Seedance 2.0 Mini reference-to-video — character-consistent generation from a portrait reference at the lowest cost in the Seedance 2.0 series.ByteDanceNewSeedance 2.0 Mini (T2V)ByteDance Seedance 2.0 Mini text-to-video — the most affordable tier of the Seedance 2.0 series. Supports 480p and 720p with native audio generation, ideal for rapid iteration and short-form content.ByteDanceNewSeedance 2.0 Mini (V2V)ByteDance Seedance 2.0 Mini video-to-video — edit or extend an existing video at the lowest cost in the Seedance 2.0 family. Ideal for high-volume batch editing workflows.Grok Imagine
xAI · 7 models
xAINewGrok Edit VideoGrok Edit Video — instruction-based video editing that replaces or transforms elements while keeping all subjects and motion intact.xAINewGrok Extend VideoGrok Extend Video — seamlessly continues an existing video clip for 4–8 additional seconds while preserving cinematic pacing and motion.xAINewGrok Imagine 1.5 (I2V)xAI Grok Imagine Video 1.5 image-to-video — animate one image into a 1–15 second cinematic video with synchronized audio at up to 1080p.xAINewGrok Imagine 1.5 (Reference)xAI Grok Imagine Video 1.5 reference-to-video — guide a 1–15 second video with up to seven image references while preserving subjects, style, and visual identity.xAINewGrok Imagine 1.5 (T2V)xAI Grok Imagine Video 1.5 text-to-video — generate 1–15 second videos with synchronized audio from a prompt at up to 1080p.xAINewGrok Reference VideoGrok Reference Video — generates video from up to 7 reference images, maintaining visual identity across shots with resolution selection.xAINewGrok Video (I2V)Grok Imagine Video image-to-video — xAI's image-to-video model with resolution selection and per-image pricing.Bernini
Krea · 4 models
KreaNewBernini-R (T2V)Bernini-R text-to-video — ByteDance's unified video model with latent semantic planning for coherent, high-quality clips.KreaNewBernini-R Edit VideoBernini-R video editing — change objects, weather, background, or camera angle with a natural-language instruction while keeping the rest of the scene intact.KreaNewBernini-R Reference EditBernini-R reference-guided video editing — garment swaps, subject insertion, and other edits guided by up to five reference images while preserving the source video.KreaNewBernini-R Reference VideoBernini-R reference-to-video — turn up to five reference images into one continuous, consistent video with smooth stable camera motion and no scene cuts.
Gemini
Google · 4 models
GoogleNewGemini Omni Flash 1.1Google Gemini Omni Flash 1.1 — GA any-to-any video generation with native synchronized audio, cinematic control, and reference-image subject consistency in 3–10 second clips up to 4K.GoogleNewGemini Omni Flash 1.1 (Edit)Google Gemini Omni Flash 1.1 conversational video editing — change elements, lighting, background, style, motion, or timing while preserving everything else. Editing uploaded videos is unavailable in the EEA, Switzerland, and the UK; editing Omni-generated videos there is supported.GoogleNewGemini Omni Flash 1.1 (Extend)Extend the end of an uploaded or previously generated Gemini Omni video with a coherent 3–10 second continuation, preserving characters, motion, and audio across turns (up to 40 seconds total).GoogleNewGemini Omni Flash 1.1 (I2V)Google Gemini Omni Flash 1.1 image-to-video — animate a first frame or interpolate between first and last frames with synchronized audio, 3–10 second duration, and output up to 4K.
Happy Horse
Alibaba · 4 models
AlibabaNewHappy Horse (Edit)Alibaba Happy Horse video-edit transforms an input video by style, weather, or setting while preserving the original motion.AlibabaNewHappy Horse 1.1 (I2V)Alibaba Happy Horse 1.1 image-to-video animates a start frame with cinematic motion, supporting durations from 3 to 15 seconds.AlibabaNewHappy Horse 1.1 (Reference)Alibaba Happy Horse 1.1 reference-to-video places a reference character naturally into any prompted scene with cinematic flair and identity consistency.AlibabaNewHappy Horse 1.1 (T2V)Alibaba Happy Horse 1.1 text-to-video generates cinematic clips from a prompt, with broad aspect ratio and resolution support.
HeyGen
HeyGen · 4 models
HeyGenNewHeyGen Avatar V Digital TwinHeyGen Avatar V Digital Twin — studio-quality talking-avatar video from your Digital Twin, driven by a script + voice. Routed directly to HeyGen's v3 API with the Avatar V engine.HeyGenNewHeyGen Lipsync PrecisionHeyGen Lipsync Precision delivers ultra-precise phoneme matching for professional dubbing, resyncing every mouth shape with maximum fidelity.HeyGenNewHeyGen Lipsync SpeedHeyGen Lipsync Speed re-dubs any speaking video with a new audio track quickly and naturally.HeyGenNewHeyGen Video AgentHeyGen Video Agent generates cinematic talking-head videos from a portrait image and an audio clip or text prompt, with realistic facial motion and atmosphere.
Hunyuan Video
Tencent · 4 models
TencentHunyuan 1.5Tencent Hunyuan 1.5 text-to-video — full diffusion workbench with inference steps, guidance, and frame count controls for high-quality video generation.TencentHunyuan 1.5 (i2v)Tencent Hunyuan 1.5 image-to-video — animate a start image with full diffusion controls including inference steps, guidance, and frame count.TencentNewHunyuan V2VTencent Hunyuan video-to-video style transfer — restyle existing footage with a text prompt while preserving motion structure. Pro mode doubles inference steps for higher quality.TencentNewHunyuan Video FoleyTencent Hunyuan Foley — generate cinematic audio tightly synced to on-screen action from a silent video, guided by a text description of the desired sounds.
Mirelo
Mirelo · 4 models
MireloNewMirelo SFX (Video → Audio)Mirelo SFX v1 video-to-audio — analyses visuals and returns a matching audio file you can layer onto the original video.MireloMirelo SFX (Video → Video)Mirelo SFX v1 video-to-video — analyses the visuals and generates matching sound effects baked directly into the output video.MireloNewMirelo SFX v1.5 (Video → Audio)Mirelo SFX v1.5 video-to-audio — improved generation quality with a start-offset option for precise audio timing control.MireloNewMirelo SFX v1.6 (Video → Video)Mirelo SFX v1.6 video-to-video — richer, higher-fidelity sound effects baked into output video with extended duration support via sliding-window generation.
AI Avatar
fal.ai · 3 models
fal.aiNewAI Avatar Multi (Audio)AI Avatar Multi generates a two-person conversation video from a portrait photo and audio files, with synchronized lip-sync for both speakers.fal.aiNewAI Avatar Multi (Text)AI Avatar Multi Text generates a two-person talking video from a portrait photo and typed dialogue — each speaker gets a separate voice.fal.aiNewAI Avatar Text (MultiTalk)AI Avatar generates a talking video from a portrait photo and typed text — pick a voice and the model handles TTS, lip sync, and natural expressions.
VEED
VEED · 3 models
VEEDNewVEED Fabric 1.0VEED Fabric 1.0 produces precise AI lip sync from a portrait image and audio clip, perfectly matching every word of the audio track.VEEDNewVEED Lipsync 2.0VEED Lipsync 2.0 replaces speech in an existing human video with production-quality lip sync, preserving the source resolution while transferring timing, emotion, and speaking style from the new audio.VEEDNewVEED SubtitlesVEED Subtitles burns professional animated captions into any video with 28 style presets, 100+ languages, and 98%+ transcription accuracy.
Bria
Bria · 2 models
BriaNewBria Green Screen DespillRemove green-screen color spill from a keyed video while preserving edges and the original audio track.BriaNewBria VRMBG 3.0Bria VRMBG 3.0 removes the background frame-perfectly from any video. Choose a replacement color or keep it transparent, preserve the original audio, and pick your output codec.
CogVideoX
CogVideoX · 2 models
Luma
Luma · 2 models
Aurora
Aurora · 1 model
Auto
Auto · 1 model
CassetteAI
CassetteAI · 1 model
DreamActor
DreamActor · 1 model
FlashHead
FlashHead · 1 model
LatentSync
LatentSync · 1 model
MMAudio
MMAudio · 1 model
Omnihuman
Omnihuman · 1 model
RIFE
RIFE · 1 model
SAM Audio
Meta · 1 model
SeedVR
ByteDance · 1 model
Sync
Sync · 1 model
ThinkSound
ThinkSound · 1 model
Trim
Trim · 1 model