AI video models, compared

MooseAI puts 47 AI video models in one studio, in your browser: Google Veo and Gemini Omni, Kling, Seedance, Hailuo, Wan, Vidu and more. Here is what each one can do. Open any of them to make a video from a description or bring a photo to life.

Open the studio

Explore video models

View all 47 models →

Which model should I use?

Each video uses credits, and the studio shows how many before you generate.

All models side by side

ModelMade byFrom textFrom a photoSound switchLongestTop resolution
VEO 3.1GoogleYesYes—8s—
VEO 3.1 FastGoogleYesYes—8s—
VEO 3.1 LiteGoogleYesYes—8s—
Gemini Omni FlashGoogleYesYes—10s—
Gemini Omni 1.1 FlashGoogleYesYes—10s4K
Gemini Omni 1.2 FlashGoogleYesYes—15s4K
Kling V2.5 ProKuaishouYesYes—10s—
Kling V2.6 ProKuaishouYesYesYes10s—
Kling V3 ProKuaishouYesYesYes15s—
Kling V3 StandardKuaishouYesYesYes15s—
Kling O1KuaishouYesYes—10s—
Kling 4.0KuaishouYesYes—30s1080p
Kling 4.0 FlashKuaishouYesYes—20s—
Seedance 2.0ByteDanceYesYesYes15s4K
Seedance 2.0 FastByteDanceYesYesYes15s720p
Seedance 2.0 MiniByteDanceYesYesYes15s720p
Seedance 2.5ByteDanceYesYesYes30s720p
Seedance 2.5 FlashByteDanceYesYesYes30s720p
MiniMax H3MiniMaxYesYes—15s4K
MiniMax H3 MaxMiniMaxYesYes—15s1080p
MiniMax H3 Max TurboMiniMaxYesYes—15s1080p
Grok Video 1.0xAIYesYes—15s—
Grok Video 1.5xAINoYes—15s720p
Grok Video 1.5 LitexAIYesYesYes15s1080p
Vidu Q3 ProShengshu TechnologyYesYesYes16s—
Vidu Q4Shengshu TechnologyNoYes—16s4K
Vidu Q3 TurboShengshu TechnologyYesYesYes16s—
Wan v2.2 FastAlibabaYesYes———
Wan 3.0AlibabaYesYesYes30s1080p
Wan 3.0 PrimeAlibabaYesYesYes30s1080p
Happy HorseAlibabaYesYes—15s1080p
Pixverse V4.5PixVerseYesYes—8s—
Pixverse V5PixVerseYesYes—8s—
Pixverse V5.6PixVerseYesYesYes10s—
Pixverse V6PixVerseYesYesYes15s1080p
Ray Flash 2 540pLuma AIYesYes———
Ray Flash 2 720pLuma AIYesYes———
Ray 2 540pLuma AIYesYes———
Ray 2 720pLuma AIYesYes———
Runway Gen 4.5RunwayYesYes—10s—
LTX 2.3 FastLightricksYesYes—20s4K
LTX 2.5 FastLightricksYesYesYes20s4K
LTX 2.3 ProLightricksYesYes—10s4K
LTX 2 TurboLightricksYesYes—10s—
FLUX 3 VideoBlack Forest LabsYesYesYes20s1080p
P-Video 2 ProPruna AIYesYes—15s4K
P-Video 2 Pro AvatarPruna AINoYes——4K

A dash means the studio has no setting for it, and the model decides.

The model families

Google Veo: VEO 3.1, VEO 3.1 Fast, VEO 3.1 Lite

Veo is Google DeepMind’s family of video models. Veo 3.1, released in October 2025, follows prompts closely and can start from your image, end on a chosen last frame or keep a look from reference images. Veo 3.1 Fast and Veo 3.1 Lite are quicker, lower-cost versions.

Gemini Omni: Gemini Omni Flash, Gemini Omni 1.1 Flash, Gemini Omni 1.2 Flash

Gemini Omni Flash is the video model of Google’s Gemini Omni family, introduced at Google I/O in May 2026. It is built to keep characters, scenes and motion consistent, and makes clips with sound. Gemini Omni 1.1 Flash, from August 2026, adds output up to 4K. Gemini Omni 1.2 Flash, from November 2026, makes clips of up to 15 seconds and holds faces, clothing and product details more closely across a clip.

Kling: Kling V2.5 Pro, Kling V2.6 Pro, Kling V3 Pro, Kling V3 Standard, Kling O1, Kling 4.0, Kling 4.0 Flash

Kling is Kuaishou’s family of video models, known for realistic motion and camera work. Kling 2 comes in Standard, Pro and Master versions; Kling 3.0, from February 2026, makes clips of up to 15 seconds with several shots and sound generated with the picture. Kling 4.0 makes clips of 3 to 30 seconds at up to 1080p, from a description, start and end frames, or reference images and characters.

Seedance: Seedance 2.0, Seedance 2.0 Fast, Seedance 2.0 Mini, Seedance 2.5, Seedance 2.5 Flash

Seedance is ByteDance’s family of video models, known for smooth, stable motion and for following instructions closely. Seedance 1.5 Pro added sound generated with the picture; Seedance 2.0, from February 2026, makes video and audio together; Seedance 2.5 makes clips of up to 30 seconds in one take.

MiniMax Hailuo: MiniMax H3, MiniMax H3 Max, MiniMax H3 Max Turbo

MiniMax makes the Hailuo video models, known for lifelike physics and expressive people. Hailuo 02 came out in June 2025 and Hailuo 2.3 in October 2025; MiniMax H3, from July 2026, is the third generation, with clips of up to 15 seconds.

Grok Imagine: Grok Video 1.0, Grok Video 1.5, Grok Video 1.5 Lite

Grok Imagine is xAI’s image and video generator, the one in Grok. Grok Imagine Video 1.5, from mid-2026, animates a still image into a clip of up to 15 seconds with sound, keeping the image’s details and lighting.

Vidu: Vidu Q3 Pro, Vidu Q4, Vidu Q3 Turbo

Vidu is Shengshu Technology’s family of video models. Vidu Q3, from January 2026, makes clips of up to 16 seconds with sound and can cut between shots; Vidu Q4, from October 2026, animates an image at up to 4K.

Wan: Wan v2.2 Fast, Wan 3.0, Wan 3.0 Prime

Wan is Alibaba’s family of video models, from its Tongyi Lab. Wan 2.1 and 2.2 were released with open weights; Wan 2.5 to 2.7 added sound and longer clips; Wan 3.0, from August 2026, makes clips of 2 to 30 seconds with a matching audio track.

Happy Horse: Happy Horse

Happy Horse is an Alibaba video model that appeared anonymously in April 2026 and went straight to the top of the Artificial Analysis video arena before Alibaba said it had made it. It makes 1080p video and its sound together.

PixVerse: Pixverse V4.5, Pixverse V5, Pixverse V5.6, Pixverse V6

PixVerse makes fast video models popular for social clips, with styles and effects. PixVerse V6, from March 2026, makes clips with sound and several shots from one prompt.

Luma Ray: Ray Flash 2 540p, Ray Flash 2 720p, Ray 2 540p, Ray 2 720p

Ray 2 is Luma AI’s video model, trained on video to give natural, coherent motion. Ray Flash 2 is a faster version at about a third of the cost; each comes at 540p or 720p.

Runway: Runway Gen 4.5

Gen-4.5 is Runway’s video model, released in December 2025. At launch it ranked first on the Artificial Analysis Video Arena, with better physics, motion and prompt following than Gen-4.

LTX: LTX 2.3 Fast, LTX 2.5 Fast, LTX 2.3 Pro, LTX 2 Turbo

LTX is Lightricks’ family of video models. LTX-2 was released with open weights and makes video with synchronized sound at up to 4K; Fast and Pro versions trade speed for quality.

FLUX: FLUX 3 Video

FLUX 3 is Black Forest Labs’ first video model, from the makers of the FLUX image models, announced in July 2026. It makes clips of up to 20 seconds with dialogue, sound effects and ambient sound generated with the picture.

Pruna P-Video: P-Video 2 Pro, P-Video 2 Pro Avatar

Pruna AI works on making AI models faster and cheaper to run, and offers its own image and video models through its API. P-Video 2 Pro makes clips of 5 to 15 seconds at up to 4K from a text prompt, start and end frames or reference images, in a speed, quality or lower-cost mode. P-Video 2 Pro Avatar makes a portrait speak or sing an audio file, lip-synced, for as long as the audio runs, at up to 4K.