The right model for every job
38 models for video, image, audio and 3D from Kuaishou, ByteDance, MiniMax, Alibaba, Google, OpenAI, ElevenLabs, Meshy, Tripo, Tencent, Hyper3D, Microsoft Research and xAI, in one desk; Auto picks for you. Prices are for the smallest job; the desk shows the exact cost before you generate.
Model families: Seedream AINano Banana AI
Compared: Kling 4.0 vs Seedance 2.5Nano Banana Pro vs 2.1Seedream vs SeedanceAI video prices
Video models
Kling 4.0
Kuaishou · Image to video · Text to video
NewKling's newest model: the steadiest motion of the family, in 720p or 1080p.
From 32 cr · 2–5 minutes · Try →
Gemini Omni Flash 1.2
Google · Image to video · Text to video
NewGoogle's Gemini Omni Flash 1.2: a photo becomes a clip with its own sound.
From 17 cr · 1–3 minutes · Try →
Kling 3.0 Pro
Kuaishou · Image to video · Text to video
NewKling 3.0 Pro: cinematic camera language and faces that hold, at 1080p.
From 23 cr · 1–3 minutes · Try →
Kling 2.6 Pro
Kuaishou · Image to video · Text to video
Kling's 1080p model: the most convincing depth and light for hero shots.
From 14 cr · 1–3 minutes · Try →
Kling 2.5 Turbo
Kuaishou · Image to video
Cinematic motion and natural parallax; a strong all-rounder.
From 9 cr · 1–2 minutes · Try →
Seedance 2.5
ByteDance · Image to video
NewestByteDance's newest Seedance, hosted in the US: the most detailed motion in the family.
From 119 cr · 2–4 minutes · Try →
Seedance 2.0 Fast
ByteDance · Image to video · Text to video
NewByteDance's Seedance 2.0 in its fast tier: fluid, realistic motion in 720p.
From 61 cr · 1–2 minutes · Try →
Seedance 1.0 Pro
ByteDance · Image to video · Text to video
Quick, steady AI camera moves with real depth at the lowest AI price.
From 5 cr · About 1 minute · Try →
MiniMax H3
MiniMax · Image to video
NewHailuo's successor: expressive people and natural physics; 768p, or 2K upscaled.
From 12 cr · 1–3 minutes · Try →
Hailuo 2.3
MiniMax · Image to video
Realistic physics and light — water, foliage and curtains move naturally.
From 12 cr · 1–3 minutes · Try →
Wan 3.0 Prime
Alibaba · Image to video · Text to video
NewAlibaba's Wan 3.0 Prime: close prompt following and clean 1080p.
From 28 cr · 1–3 minutes · Try →
HappyHorse 1.1
Alibaba · Image to video · Text to video
NewAlibaba's HappyHorse 1.1: lively motion from one photo, with 1080p at a low price.
From 28 cr · 1–3 minutes · Try →
Grok Imagine 1.5
xAI · Image to video · Text to video
NewxAI's Grok Imagine 1.5: expressive, stylised motion from one photo.
From 34 cr · 1–3 minutes · Try →
Airift Motion
Airift · Image to video
FreeA smooth virtual camera move across your photo. Nothing in the picture changes.
Free · About 10 seconds · Try →
Image models
Seedream 5.5
ByteDance · Image edit · Text to image
NewestByteDance's newest image model: it changes what you ask for and leaves the rest of the picture as it was.
From 1 cr · Under a minute · Try →
Kling Image 4.0
Kuaishou · Image edit · Text to image
NewKling's image model: edits a photo from an instruction, at 2K or 4K.
From 3 cr · Under a minute · Try →
Nano Banana Pro
Google · Image edit · Text to image
NewGoogle's Pro editor: the most faithful edits, text and fine detail intact, up to 4K.
From 6 cr · Under a minute · Try →
Nano Banana 2.1
Google · Image edit · Text to image
NewGoogle's newest everyday editor: quick, inexpensive, keeps the picture as it is.
From 2 cr · Under a minute · Try →
Nano Banana 2
Google · Image edit · Text to image
Gemini's image editor: keeps shapes, lines and perspective most consistently.
From 4 cr · Under a minute · Try →
Nano Banana Lite
Google · Image edit · Text to image
NewGoogle's efficiency editor: a small change back in about ten seconds, at 1K.
From 2 cr · About 10 seconds · Try →
Nano Banana 1
Google · Image edit · Text to image
The first Nano Banana, Gemini 2.5 Flash Image: the editor that started the name, at 1K.
From 2 cr · Under a minute · Try →
GPT Image 2
OpenAI · Image edit · Text to image
NewOpenAI's image model: strong instruction following for restyling.
From 9 cr · About a minute · Try →
Seedream 4.5
ByteDance · Image edit · Text to image
Fast, inexpensive edits with vivid, magazine-style styling.
From 2 cr · Under a minute · Try →
Qwen Image 3.0
Alibaba · Image edit · Text to image
NewAlibaba's newest image model: edits a photo from an instruction, in Chinese or English.
From 2 cr · Under a minute · Try →
Grok Imagine Image 2.0
xAI · Image edit · Text to image
NewxAI's image model: bold, stylised edits of a photo from an instruction.
From 3 cr · Under a minute · Try →
Qwen Image
Alibaba · Image edit · Text to image
Precise instruction-following that returns the photo at its original size.
From 1 cr · Under a minute · Try →
Audio models
Eleven v4
ElevenLabs · Text to speech
NewElevenLabs' newest voices: expressive, studio-quality speech that follows the punctuation.
From 4 cr · About 10 seconds · Try →
MiniMax Speech 2.8 HD
MiniMax · Text to speech
MiniMax's HD voices: natural, even narration with named character voices.
From 4 cr · About 10 seconds · Try →
Gemini 3.8 Flash TTS
Google · Text to speech
NewGoogle's speech model: 30 voices and the lowest price per character.
From 2 cr · About 10 seconds · Try →
Lyria 3.5
Google · Music
NewGoogle DeepMind's music model: a finished three-minute track from a sentence, with or without vocals.
From 4 cr · About a minute · Try →
ElevenLabs Music 2.5
ElevenLabs · Music
ElevenLabs' music model: polished, radio-ready tracks at the length you choose.
From 24 cr · 1–2 minutes · Try →
MiniMax Music 3
MiniMax · Music
NewMiniMax's song model: sings your own lyrics, with structure tags for verse and chorus.
From 3 cr · 1–2 minutes · Try →
ElevenLabs Sound Effects v3
ElevenLabs · Sound effects
NewestAny sound from a description, up to 30 seconds, with a seamless loop option.
From 1 cr · About 20 seconds · Try →
3D models
Meshy T2.5
Meshy · Image to 3D · Text to 3D
NewestMeshy's newest Smart Topology model: clean, game-ready meshes from one photo.
From 8 cr · 3–6 minutes · Try →
Tripo P2
Tripo3D · Image to 3D · Text to 3D
Tripo's production model: detailed geometry with PBR textures from a single image.
From 40 cr · 2–5 minutes · Try →
Hunyuan 3D 3.1 Pro
Tencent · Image to 3D · Text to 3D
Tencent's Hunyuan 3D Pro: faithful, high-resolution textured models from one image.
From 15 cr · 2–4 minutes · Try →
Rodin 2.5
Hyper3D · Image to 3D
Hyper3D's Rodin: realistic, production-ready models with PBR materials.
From 16 cr · 2–5 minutes · Try →
Trellis 2
Microsoft · Image to 3D
NewMicrosoft Research's Trellis 2: a textured model from a photo in about a minute, at the lowest price.
From 12 cr · About a minute · Try →
Create AI videos from your photos, all in one place.
Free to start. No card. The AI engines by the credit.