AI models available in Percify

Every image, video, lip-sync and audio model you can run in Percify, grouped by what it makes, with clip lengths and credit prices from the Playground.

Percify gives you more than 100 AI models for video, images, lip-sync, speech and music. You can run any of them in the Playground and pay for each run in credits. The tables below list the models in the Playground catalog on 16 September 2026, grouped by what they make, with the clip lengths they offer and what a run costs.

Models are added and prices can change. The Playground at app.percify.io/playground always shows the current list, and each model’s Run button shows the exact price for your settings before you run it.

How to read the credit prices

• A single number is a flat price per run. Changing the settings does not change it.
• A range means the price depends on your settings. The low end is the shortest clip at the lowest quality or resolution. The high end is the longest clip at the highest.
• Per second prices are charged on the length of the file you upload, such as the audio for lip-sync or the video you want to edit. A minimum applies to very short files.
• A run that fails is refunded. See How Percify credits work.

Video models

Text to video

Describe a shot and get a clip back. Several of these also accept optional reference images, videos or audio in the Playground form.

Model	Clip lengths	Credits
Gemini Omni Flash Video	3 to 10 s	39 to 130
Grok Imagine (Text-to-Video)	6 or 10 s	60 to 100
Hailuo 02 Standard (Text-to-Video)	6 or 10 s	46 to 77
Hailuo 2.3	6 or 10 s	34 to 57
Hailuo 2.3 Pro	Set by the model	49
Kling V3 Turbo Pro	3 to 15 s	43 to 215
Kling V3 Turbo Standard	3 to 15 s	34 to 170
Kling V3.0 Pro	3 to 15 s	34 to 170
Kling v3 Standard (Text-to-Video)	3 to 15 s	50 to 250
Kling Video O3 Standard	3 to 15 s	26 to 130
LTX-2 Fast (Text-to-Video)	6 to 20 s, in 2 s steps	48 to 160
LTX-2 Pro	6, 8 or 10 s	36 to 60
Luma Ray 3.2	5 or 10 s	50 to 200
Ovi (Video + Audio)	Set by the model	15
P-Video (Text-to-Video)	1 to 20 s	4 to 160
Pika 2.2	5 or 10 s	20 to 40
PixVerse V5 (Text-to-Video)	5 or 8 s	30 to 129
Pixverse V6	1 to 15 s	3 to 141
Realtime Video (Text to Video)	5 to 15 s	20 to 90
Seedance 2.0 Fast Text-to-Video	5, 10 or 15 s	100 to 900
Seedance 2.0 Mini	4 to 15 s	30 to 900
Seedance 2.0 Text-to-Video	5, 10 or 15 s	100 to 900
Seedance 2.5 Text-to-Video	4 to 30 s	96 to 7,200
Seedance 2.5 Text-to-Video Turbo	4 to 30 s	120 to 991
Seedance V1.5 Pro (Text-to-Video)	4 to 12 s	42 to 126
Sora 2 (Text-to-Video)	4, 8, 12, 16 or 20 s	80 to 400
Sora 2 Pro	4, 8, 12, 16 or 20 s	120 to 1,404
Veo 3 Fast (Text-to-Video)	4, 6 or 8 s	240
Veo 3.1	4, 6 or 8 s	160 to 320
Veo 3.1 Fast	4, 6 or 8 s	60 to 120
Veo 3.1 Lite (Text-to-Video)	4, 6 or 8 s	80 to 128
Wan 2.2 Ultra-Fast 480p (Text-to-Video)	5 or 8 s	10 to 16
Wan 2.5 (Text-to-Video)	5 or 10 s	50 to 300
Wan 2.7	2 to 15 s	50 to 225
Wan 3.0 Reference-to-Video	2 to 30 s	16 to 960
Wan 3.0 Text-to-Video	2 to 30 s	16 to 960

Image to video

Start from a still image and describe the motion. Some models also take an optional last frame.

Model	Clip lengths	Credits
Hailuo 2.3 (Image to Video)	6 or 10 s	34 to 57
Kling V3 Turbo Pro (Image to Video)	3 to 15 s	43 to 215
Kling V3 Turbo Standard (Image to Video)	3 to 15 s	34 to 170
Leonardo Motion 2.0	Set by the model	30
LTX-2 Pro (Image to Video)	6, 8 or 10 s	36 to 60
Luma Ray 3.2 (Image to Video)	5 s	50 to 200
P-Video Turbo (also asks for an audio file)	5 or 10 s	10 to 20
Pika 2.2 (Image to Video)	5 or 10 s	20 to 40
Pixverse V6 (Image to Video)	1 to 15 s	3 to 141
Realtime Video (Image to Video)	5 to 15 s	20 to 90
Seedance 2.0 Fast Image-to-Video	5, 10 or 15 s	100 to 900
Seedance 2.0 Image-to-Video	5, 10 or 15 s	100 to 900
Seedance 2.0 Mini (Image to Video)	4 to 15 s	30 to 900
Seedance 2.5 Image-to-Video	4 to 30 s	96 to 7,200
Seedance 2.5 Image-to-Video Turbo	4 to 30 s	120 to 991
Seedance V1.5 Pro I2V	5 or 10 s	25 to 250
SkyReels V4 (Image to Video)	3 to 15 s	55 to 527
Sora 2 I2V	4, 8 or 12 s	70 to 210
Sora 2 Pro I2V	4, 8 or 12 s	90 to 405
Veo 3.1 (Image to Video)	4, 6 or 8 s	160 to 320
Veo 3.1 Fast (Image to Video)	4, 6 or 8 s	60 to 120
Wan 2.7 (Image to Video)	2 to 15 s	50 to 225
Wan 2.7 Pro (Image to Video)	5, 10 or 15 s	60 to 242
Wan 3.0 Image-to-Video	2 to 30 s	16 to 960

Lip-sync and talking video

Give a face image and an audio file, and the face speaks the audio.

Model	You give it	Credits
InfiniteTalk Fast	A face image and audio	2 per second of audio, minimum 4
InfiniteTalk	A face image and audio	4 per second at 480p, 6 per second at 720p, minimum 8

These two models also power Fast and High mode in Clone Yourself. See Lip-sync.

Motion control, character swap and video editing

Model	You give it	Credits
Grok Imagine Video Edit	A prompt, with an optional image or video to edit	6 to 90 (6 per second, 1 to 15 s)
Kling v2.6 Standard Motion Control	A character image, a motion video and a prompt	5 per second of video, minimum 15
Kling v3 Motion Control	A character image and a motion video	9 per second of video, minimum 27
Lucy Edit Pro (Video Edit)	A video and a prompt	7 per second at 480p, 10.5 per second at 720p, minimum 7
MoCha Character Swap	A video and a reference image	3 per second at 480p, 6 per second at 720p, minimum 15
P-Video Replace	A video and 1 to 3 reference images	2.1 per second at 720p, 4.2 per second at 1080p, minimum 3
Wan 2.2 Animate Replace	A video and a new character image	3 per second at 480, 6 per second at 720, minimum 15
Wan 2.7 Video Edit	A video, a prompt and up to 3 optional reference images	14 per second at 720p, 21 per second at 1080p, minimum 28

Dubbing and video translation

Model	You give it	Credits
ElevenLabs Dubbing	A video or audio file and a target language (18 to choose from)	72 per minute of media, minimum 72
HeyGen Video Translate	A video and an output language (178 options)	3 per second of video, minimum 90

See also Dubbing.

Image models

Text to image

Model	Input	Credits
Flux 2 Dev	Prompt, optional reference images	5
Flux 2 Klein 4B	Prompt, optional reference images	5
Flux Kontext Pro	Prompt, optional image to edit	5
Flux Schnell	Prompt	2
Gemini 3 Pro Image	Prompt	25
GPT Image 2	Prompt, optional reference images	4 (low), 8 (medium), 16 (high)
GPT Image 2.5 Flare	Prompt	1 to 80, by quality and resolution (1K, 2K, 4K)
GPT Image 2.5 Sunburst	Prompt	1 to 80, by quality and resolution (1K, 2K, 4K)
Grok Imagine Image	Prompt, optional image	5
Ideogram V3 Quality	Prompt, optional image, mask and reference images	9
Ideogram V4	Prompt, optional image	10
Image 01	Prompt	1
Imagen 4	Prompt	4
Imagen 4 Ultra	Prompt	6
Kling Image V3	Prompt	3
Luma Photon	Prompt	2
MAI Image 2.5	Prompt	5
Midjourney	Prompt	10
Nano Banana Pro	Prompt	25
Qwen Image 2.0	Prompt	3
Recraft V4	Prompt	4
Recraft V4.1 Pro	Prompt	25
Runway Gen-4 Image	Prompt, optional reference images	8
SeedReam 5 Lite	Prompt, optional image	5
Seedream 5.0 Pro	Prompt	9
Z-Image Turbo	Prompt	2

Image editing

Model	You give it	Credits
Bria GenFill (Inpaint)	An image, a mask and a prompt	4
Flux Kontext Fast	An image and a prompt	5
GPT Image 2.5 Flare Edit	Up to 16 images and a prompt	2 to 82 by quality and resolution, plus 2 for each image after the first
GPT Image 2.5 Sunburst Edit	Up to 16 images and a prompt	2 to 82 by quality and resolution, plus 2 for each image after the first
Nano Banana 2 (Edit)	1 to 14 images and a prompt	10
NVIDIA ChronoEdit	An image and a prompt	2
Qwen Image 2.0 Edit	Images and a prompt	3
Seedream 5.0 Lite Edit	Images and a prompt	4
Wan 2.7 Image Edit Pro	Images and a prompt	8
Z-Image Turbo Img2Img	An image and a prompt	5

Faces, portraits and text in images

Model	You give it	Credits
Image Head Swap	A base image and a face photo. It keeps the body, pose and background.	10
Infinite You	A face photo and a scene description	10
Nano Banana 2 (Edit Fast)	Your photo, a reference image and a prompt	10
AI Image Translator	An image and a target language	12
Qwen Image Translate	An image and a target language	2

Audio models

Text to speech and voice cloning

Model	Input	Credits
Chatterbox Turbo	Text (up to 500 characters), a preset voice or an optional voice sample longer than 5 seconds	5
ElevenLabs Eleven V3	Text and a voice	10
ElevenLabs Multilingual V2	Text and a voice	10
Gemini TTS	Text, a voice and optional style instructions	Varies with script length
Qwen3 TTS Flash	Text and a voice	2
Seed Speech TTS 2.0	Text, a preset voice and an optional delivery instruction	3
Speech 2.8 HD	Text and a voice	10
Speech-02-HD	Text and a voice	5
Speech-02-Turbo	Text and a voice	5
XTTS-v2	A voice sample, text and an output language (16 options)	10
Zonos 2	A voice sample to clone and the text to say	2 per minute of voice sample, minimum 2

To clone your own voice in Voice Studio, see Voice cloning.

Music and sound effects

Model	Input	Credits
ElevenLabs Music	A description, length 5 seconds to 5 minutes, instrumental or with vocals	9
Lyria 3 Pro	A prompt, with an optional image	8
Mirelo SFX 1.6	A sound description, 0.1 to 60 seconds	10
Mureka V9 (Song)	Lyrics and an optional style prompt	5
Music 2.6	A prompt and lyrics	15
Seed Audio 1.0	A prompt, with optional reference audio or image	30

FAQ

Do I need a paid plan to use these models?

No plan unlocks or hides a model. Every model runs on credits, and the plans differ in how many credits you get each month. See How Percify credits work and Plans and payments.

Can I call these models from my own code?

Yes. The Percify API runs the models in the public catalog with an API key, and charges the same credits. API access is included on the Scale and Ultra plans. See the API reference and API authentication.

Which model should I start with?

Use Compare to give two models the same prompt and see which result you prefer before you spend credits on longer or higher-resolution clips.

Related

Playground and Compare

Run a model, or test two side by side

How to write prompts

Get better results from the prompt fields these models use

How Percify credits work

Monthly credits by plan and how runs are charged

Percify API

Run models over HTTP