Image Models
64 models available
Nano Banana Pro
State-of-the-art image generation and editing — Gemini-class
Kling O1
Kling O1 reasoning model with enhanced cinematic quality
Nano Banana 2.1
Nano Banana 2.1 — Google's newest Gemini image model. Generate from text or edit with up to 14 reference images, at 1K to 4K, with sharp text and extreme ratios from 8:1 to 1:8.
GPT-Image 1.5
GPT Image 1.5 generates high-fidelity images with strong prompt adherence, preserving composition, lighting, and fine-grained detail
Nano Banana 2
Gemini-class image generation and editing — the full-size Nano Banana 2.
GPT-Image 2.0
Next-gen image model — near-perfect text rendering (>99%), enhanced photorealism, UI/screenshot-grade detail, and superior multilingual text (JP/KR/HI/BN). Supports optional reference-image-guided generation.
GPT-Image 2.0 Edit
Precise single-image editing with detail preservation. Built on GPT-Image 2.0's natively multimodal architecture.
FLUX 3 Image
FLUX 3 Image — Black Forest Labs' newest image model. Generate in native 2K and 4K with close control over layout and clean text, or edit a photo and combine up to 10 reference images while everything you didn't change stays as it was.
FLUX Kontext Pro
Advanced image editing and transformation
Krea 3
Krea 3 — Krea's newest image model. Generate from text, or edit and combine up to 8 reference images with plain instructions: change the setting while your subject stays consistent. Quality and Fast modes, at 1K or 2K.
GPT Image 2.6 Sunburst
OpenAI GPT Image 2.6 Sunburst: the precision model for premium visual work, with extra fidelity on intricate detail and text. Choose quality from Low to Max, exact dimensions, output format and transparent backgrounds, or make an animated GIF.
GPT Image 2.6 Sunburst Edit
OpenAI GPT Image 2.6 Sunburst image editing with up to 16 source images: the tightest control, with subjects and composition kept across many rounds of edits. Medium to Max quality, or animate your image as a GIF.
MiniMax H3 Image
MiniMax H3 Image — MiniMax's image model. Generate from text at 2K or 4K in eight aspect ratios, or edit and combine up to 16 reference images with plain instructions (call them Image 1, Image 2 and so on). Optional prompt expansion turns a short idea into a detailed description.
Seedream 5.5
Seedream 5.5 — ByteDance's newest Seedream image model. Generate from text at 1K, 1.5K or 2K in eight aspect ratios, edit and combine up to 10 reference images with plain instructions, or make PNGs with a transparent background.
MiniMax H3 Image Highspeed
MiniMax H3 Image Highspeed — the Highspeed version of MiniMax H3 Image. The same controls: text-to-image at 2K or 4K in eight aspect ratios, or edit and combine up to 16 reference images with plain instructions.
Cosmos 3 Super
NVIDIA's flagship photorealistic text-to-image model with optional agentic refinement — generates multiple candidate images per round and rewrites the prompt between rounds for stronger prompt adherence. Top-end physics-aware photorealism, structured-JSON prompt expansion supported.
Seedream 5.0 Pro
Premium image model for precise text-to-image and multi-reference editing — product visuals, design assets, dense layouts, and commercial creative. Edit mode takes up to 10 reference images.
Topaz Precision Upscale
Topaz Labs Gigapixel precision image upscaling — enlarges photos faithfully up to 4x while keeping edges crisp and detail clean.
Topaz Wonder Upscale
Topaz Labs' generative Wonder 3.5 image upscaler — recovers and rebuilds detail in low-quality or damaged images while enlarging up to 4x.
GPT Image 2.5 Flare
GPT-Image 2.5 Flare — OpenAI's default image model introduced with ChatGPT Images 2.5. Fast, high-quality generation with natural lighting, rich textures, complex layouts and transparent backgrounds (PNG/WebP).
GPT Image 2.5 Flare Edit
GPT-Image 2.5 Flare for image editing — describe the change to a source image, with up to 15 extra reference images. Supports transparent output.
GPT Image 2.5 Sunburst
GPT-Image 2.5 Sunburst — OpenAI's precision image model for premium visual work. Extra fidelity on intricate detail in exchange for longer generation times.
GPT Image 2.5 Sunburst Edit
GPT-Image 2.5 Sunburst for image editing — precise, detail-preserving edits from an instruction and a source image, plus up to 15 references.
Ideogram 4.5
Ideogram 4.5 — Ideogram's newest image model, with sharp, accurate text in images. Generate from a prompt in 15 aspect ratios at low, medium or high quality, or edit a photo with up to four reference images.
GPT Image 2.6 Flare
OpenAI GPT Image 2.6 Flare: fast, high-quality image generation with natural lighting, rich textures and complex layouts. Choose quality from Low to Max, exact dimensions, output format and transparent backgrounds, or make an animated GIF.
GPT Image 2.6 Flare Edit
OpenAI GPT Image 2.6 Flare image editing with up to 16 source images: changes only what you ask and keeps the subject, composition and background. Medium to Max quality, or animate your image as a GIF.
Meta Muse Image
Meta's Muse Image model has faithful instruction-following and exceptional visual fidelity, with fine details like text, plots, and QR codes rendered accurately. Generates images from text prompts in nine aspect ratios at a flat $0.01 per image.
Meta Muse Image Edit
Meta's Muse Image model in editing mode — precise edits that change only what you ask, stay coherent across turns, and compose from multiple reference images. Accepts 1-10 reference images at $0.01 per output image.
Wan 2.7 Image Pro
Wan 2.7 image model, higher quality tier — unified text-to-image and image editing.
Ideogram V4
Ideogram's V4 — best-in-class in-image text rendering, sharper typography, stronger prompt adherence, and three rendering speeds (TURBO / BALANCED / QUALITY). Native multi-aspect support, no fine-tuning needed for posters, logos, thumbnails, ads.
Seedream 5.0 Pro Layer Decomposition
Split an image into editable layers — background, subject, and each key visual element as its own transparent PNG. Built on Seedream 5.0 Pro's image understanding, so structure and detail survive the separation. The number of layers depends on the image.
Grok Imagine Image 2.0
xAI's Grok Imagine Image 2.0 — fast text-to-image with a 2K option, a low/medium quality dial, and thirteen aspect ratios from ultrawide to ultratall. Up to four images per request.
Grok Imagine Image 2.0 Fast
xAI's Grok Imagine Image 2.0 at a flat low price — sharp typography and precise prompt following, in five aspect ratios. One image per request, no quality or resolution dials.
Grok Imagine Image 2.0 Edit
xAI's Grok Imagine Image 2.0 in editing mode — hand it up to three images and it edits, composes or restyles them while holding identity and style. Same 1K/2K and low/medium tiers as the generator, up to four results per request.
Google Virtual Try-On
Google's virtual try-on model. Combines a person photo with a clothing product image into a realistic try-on shot.
Topaz Bloom Creative Upscale
Topaz Labs' Bloom 2 creative image upscaler — reinvents detail for striking enhancement. Best for AI-generated images that deserve a premium finish.
Ideogram Vector
Ideogram Vector — Ideogram's vector model. Generate an editable SVG from a prompt, guided by up to five reference images, or convert an existing image into an SVG. Every result comes as a PNG preview plus the SVG file.
Flixly Upscale
Flixly Upscale — sharpens an image at its original size or enlarges it 2x, 4x or 8x, up to about 8K. Detail and texture are restored while text and logos stay exactly as written. Takes JPG, PNG, WebP or GIF.
FLUX 3 Klein LoRA
FLUX 3 Klein LoRA — Black Forest Labs' FLUX 3 Klein base model with up to 3 of your own LoRAs. Generate images in a trained style, character or product look at up to 4K, or edit and combine up to 10 reference images with it.
Krea 2 Turbo
Krea 2 Turbo — speed-optimized text-to-image generation with the full aesthetic range of Krea 2. Fast ideation with prompt expansion and acceleration modes.
MAI-Image-2.5
Microsoft's MAI-Image-2.5 — general-purpose text-to-image with strong photorealism, prompt adherence, and in-image text rendering. Supports up to 4 images per request and seven aspect ratios.
Grok Imagine — Image
Grok Imagine image variant. Text-to-image plus image-to-image variation editing with strong prompt adherence. Sibling of Grok Imagine Video.
SeedVR2 Upscale
SeedVR2 image upscaling — fast, inexpensive enlargement up to 4x. Great value for everyday upscales.
Marigold V2
Marigold V2: turn a photo into a depth map, surface normals or albedo. One image in, one map out, same size.
Wan 2.7 Image
Wan 2.7 image model — unified text-to-image and image editing.
Recraft V4.1 Flash
Recraft V4.1 Flash — Recraft's fastest text-to-image model, built for quick iteration: test a composition, adjust the prompt and try another direction in seconds. Strong on design work like logos, posters and social graphics.
Z-Image
Tongyi-MAI's lightweight 6B image model — photorealistic output, sub-second generation, and notably accurate bilingual (English + Chinese) text rendering. Apache-2.0 licensed.
Nano Banana 2 Lite
Fast, cost-efficient Gemini image model — rapid ideation, high-throughput generation, and lightweight editing. Built for concept exploration, social visuals, product drafts, and scaled production.
Image Translator
Translate the text inside an image into another language and get back a clean, localized version — great for localizing posters, menus, screenshots, and product shots.
Qwen Image 2.1
Qwen Image 2.1 — text-to-image and multi-reference editing at 1K or 2K, with native transparent backgrounds and automatic prompt enhancement. Edit mode takes up to 10 reference images.
Seedream 5.0 Flash
ByteDance's fastest Seedream 5.0 model. Generate images from text or edit with up to 10 reference images, at about 4 megapixels in eight aspect ratios.
Nano Banana
Nano Banana standard image generation
Nano Banana Edit
Nano Banana image editing
FLUX 2 Pro Edit
FLUX 2 Pro image-to-image editing
FLUX 2 Flex
FLUX 2 Flex fast text-to-image generation
Seedream 4.0
Seedream 4.0 text-to-image
Seedream 4.5
Seedream 4.5 text-to-image with improved quality
Seedream 5.0 Lite
Seedream 5.0 Lite fast text-to-image
GPT-4o Image
GPT-4o multimodal image generation
Qwen Image
Qwen text-to-image generation
Qwen 2 Image
Qwen 2 improved text-to-image
Topaz Upscale
Topaz AI image upscaling with detail enhancement
Recraft Remove BG
Recraft AI background removal
Recraft Upscale
Recraft AI crisp image upscaling
Video Models
87 models available
Veo 3.1
Premium video generation model with native audio
Kling 4.0
Kling 4.0 — Kling AI's newest video model with native sound. Text-to-video, image-to-video with an optional last frame, and reference-to-video from up to 10 images. 3 to 30 seconds at 720p or 1080p.
Seedance 2.5
ByteDance's newest video model. Generates up to 30 seconds in a single shot with synchronized audio — dialogue, music, and effects in the same pass — from a prompt, a start frame, or up to 50 image, video, and audio references.
Wan 3.0 Prime
The premium tier of Alibaba's Wan 3.0 — faster turnaround, cinematic motion, and stronger visual continuity. Text, image, and reference modes, native 30-second clips in a single pass, with audio.
MiniMax H3 Max
Post-trained MiniMax H3, ranked #1 for overall quality, prompt understanding, and aesthetics against leading video models. 5-15 second clips in native 480p, 768p or 1080p from a prompt, a start frame, or up to 9 reference images, 3 motion clips and 3 audio clips. Reference-to-video is generally available: up to 2x faster, with much stronger subject preservation.
MiniMax H3 Max Multi-Angle
MiniMax H3 Max with a built-in camera rig. Your still frame stays frozen while the camera flies a path you choose — orbit around a subject, push in, crane overhead, or set up to 12 poses by hand. 5-15 second clips in native 480p, 768p or 1080p.
Kling O3 Reference
Kling O3 Reference — Kuaishou's reference-to-video model. Give it your characters and reference images and it composes and directs the shot itself, keeping each character the same from every angle.
Kling 3.0
Latest Kling flagship with best-in-class motion and cinematic quality
Wan 3.0
Alibaba's newest video model. Native 30-second clips in a single pass — no stitching — with audio and reality-grade motion, from a prompt, a start frame, or image, video, and audio references.
MiniMax H3 LoRA
MiniMax H3 (Hailuo-03) with LoRA support for image-to-video. Animate images with custom LoRA models across four resolution tiers (480p/720p/1080p/2K), 4-15 seconds, with native stereo audio. Higher quality than base H3.
Kling 4.0 Edit
Kling AI's Kling 4.0 video editing: describe one change to a 3 to 30 second clip, such as an expression, a movement, the camera or the background, and keep the rest.
MiniMax H3
MiniMax's next-generation multimodal video model (Hailuo-03). Generates up to 2K video with native stereo audio from a text prompt or a reference image, with strong instruction following, motion, and readable on-screen text.
Gemini Omni Flash
Google's any-input video model. Create 4-10 second clips with native audio from a prompt, an image, three reference images, or an existing video — then reshape scenes with plain language. Grounded in Gemini's real-world knowledge for coherent physics and motion.
Gemini Omni Flash v1.1
Google's Omni video model v1.1 — now with resolution control (360p/720p/1080p/4K) and video editing. Create or transform 3-10 second clips with native audio from text, an image, references, or an existing video. Upgraded from v1 with full resolution flexibility.
MiniMax H3 Max Turbo
MiniMax H3 Max on a faster inference stack at half the price. The same post-trained model tuned for prompt adherence and aesthetics — 5-15 second clips from a prompt or a start frame, in native 480p, 768p or 1080p.
FLUX 3 Max
FLUX 3 Max is the fast, high-adherence version of Black Forest Labs' FLUX 3 video model. Generate 5 to 20 seconds at up to 1080p with sound, from a prompt, a photo, or a first and last frame — and get it back in a fraction of the usual wait.
Gemini Omni Flash v1.2
Google's Omni video model v1.2: 3 to 15 second clips with native audio from text, an image, or up to 10 reference images and 3 reference clips, at 360p, 720p, 1080p or 4K.
Kling 2.1 Master
Latest Kling with best-in-class motion and quality
Grok Imagine Video 1.5
Grok Imagine Video 1.5 — image-to-video with synchronized audio, strong prompt adherence, and consistent visual quality. Currently ranked #1 on the Arena blind test leaderboard for image-to-video, surpassing Seedance 2.
PixVerse V6 Max
PixVerse V6 Max — PixVerse's top V-series video model, built for film, short drama, ad and music-video work. Text-to-video or image-to-video, 1 to 15 seconds at up to 1080p, with native sound and lip-synced dialogue.
Veo 3.1 Fast
Faster, more cost-effective version of Veo 3.1
FLUX 3
Black Forest Labs' FLUX 3 — text-to-video with synchronized audio generated by the model itself. Up to 20 seconds at 1080p, no reference image required.
HeyGen Video
HeyGen Video — HeyGen's video model with its own sound, built on MiniMax H3. Text-to-video, image-to-video from a first frame, and reference-to-video that keeps the products, people or places from up to 9 images, 3 videos and 3 audio clips. 5 to 15 seconds, 480p to 2K.
Kling 2.0 Master
Kling 2.0 with significantly improved quality and motion
Wan 2.7
Wan 2.7 — text-to-video and image-to-video (first-frame / first-last-frame / continuation).
Topaz Astra Creative Upscale
Topaz Labs' Astra 2 creative video upscaler — reimagines fine detail and typically delivers 4K output. Clips up to 5 minutes.
Topaz SDR to HDR (Hyperion)
Topaz Labs' Hyperion 2.5 professional SDR-to-HDR conversion — redistributes luminance and color while preserving detail in text, faces and motion.
OmniHuman 2.0
OmniHuman 2.0 — ByteDance's talking-video model. Give it one photo of a person and a voice track of up to 5 minutes, and it makes a video of them speaking it, with lip sync, facial expressions and body movement that follow the audio. Add an optional prompt to direct the scene and camera.
Gemini Omni Flash v1.2 Edit
Google's Omni v1.2 video editing: describe a change to a clip of up to 15 seconds and keep the rest. The result is as long as your clip, at 360p to 4K.
Vidu Q4
Vidu Q4 — Shengshu's video model with native audio. Animate a first frame into a 3 to 16 second clip, or build a scene from up to 12 reference images and 3 voice clips that keeps characters and voices consistent. 540p up to 4K.
Kling 1.5 Pro
High quality video generation
Runway Gen 4.5
Newer Runway Gen 4.5 — text-to-video with optional reference image, 5/10s.
Seedance 2.0
Seedance 2.0 — text-to-video and image-to-video.
FLUX Video Upscale
Black Forest Labs' FLUX 3 powered video super-resolution up to 4K, with a source-faithful Precise mode and a Creative detail-enhancement mode. Clips up to 20 seconds.
Topaz Starlight Upscale
Topaz Labs' professional Starlight generative video upscaler (Starlight Precise 2.6) — rebuilds detail that isn't in the source, up to 4K.
MiniMax H3 Max Extend
MiniMax H3 Max for making videos longer: upload a clip, describe what happens next, and it adds 5 to 15 seconds of new footage that continues the same scene. Get the full extended video or only the new part, at 480p, 768p, 1080p or 2K.
MiniMax H3 Max Insert
MiniMax H3 Max for editing a scene into an existing video: pick where the new scene starts and where the original resumes, describe it (optionally with reference images), and get the whole video back with a 5 to 13 second generated scene in place, at 480p or 768p.
MiniMax H3 Max Recast
MiniMax H3 Max for recasting a video: upload a clip of 5 to 30 seconds and one photo per new person, and it replaces the people while keeping the original motion, camera, cuts and sound. 768p or 1080p.
VEED Dubbing Pro
VEED Dubbing Pro for video: dubs a video of up to 10 minutes into 100+ languages and regional accents such as Mexican Spanish and Brazilian Portuguese, with a closer match to each speaker's voice and lips synced to the new words.
Kling 4.0 Flash
Kling AI's fast Kling 4.0 model with native sound, for quick drafts and frequent iteration. Text-to-video, image-to-video and reference-to-video from up to 10 images, 3 to 20 seconds at 720p.
Gemini Omni Flash v1.2 Extend
Google's Omni v1.2 continues a video with 3 to 10 seconds of new footage that follows your prompt, at 360p to 4K. You pay only for the new seconds.
Sync LipSync
Professional lip sync
Kling 3.0 Motion Control
Reference-image + reference-video motion transfer. Pick whose pose drives the output via the character orientation toggle.
Kling 3.0 Turbo
Speed-optimised Kling 3.0 — faster generation at high visual quality. Text-to-video (up to 2,500-char prompts), first-frame image-to-video, and multi-shot storyboarding (up to 6 prompted shots in one clip). Standard tier is 720p, Pro tier is 1080p.
Runway Gen-4 Turbo
Runway Gen-4 Turbo video generation from text or a reference image. 5 or 10 seconds; 1080p supports 5 seconds.
MiniMax H3 Max VHS
MiniMax H3 Max with a built-in VHS camcorder look: describe the scene and sound, then pick Light, Medium or Heavy tape damage. 5-15 second clips at 768p with native audio, from a prompt or a start frame.
MiniMax H3 Max Retro Toon 70s
MiniMax H3 Max with a built-in 1970s hand-painted cartoon style, so you only write the scene, action, camera and sound. 5-15 second clips at 768p with native audio, from a prompt or a start frame (painted artwork keeps the look most consistent).
MiniMax H3 Max Low Poly
MiniMax H3 Max with a built-in retro low-poly 3D look, no style prompting needed. 5-15 second clips at 768p with native audio, from a prompt or a start frame (low-poly artwork keeps the look most consistent).
MiniMax H3 Max Hand Drawn
MiniMax H3 Max with a built-in hand-drawn pencil animation style. Describe the scene, action, camera and sound to get 5-15 second clips at 768p with native audio, from a prompt or a start frame (pencil artwork keeps the look most consistent).
MiniMax H3 Max 16-bit Pixel
MiniMax H3 Max with a built-in 16-bit pixel-art style for retro game-like scenes, no style prompting needed. 5-15 second clips at 768p with native audio, from a text prompt or a start frame (pixel artwork keeps the style most consistent).
Kandinsky 6.0 Pro
Kandinsky 6.0 Pro — Kandinsky Lab's flagship video model for cinematic, high-fidelity clips with synchronized audio. 5-second text-to-video and image-to-video at 480p, or upscaled to 1080p or 4K with its built-in super-resolution.
Kling 4.0 Flash Edit
Kling AI's fast Kling 4.0 video editing: describe a change to a 3 to 20 second clip, such as an object, a style, the weather or the background, and keep the rest. 720p.
Gemini Omni Flash v1.2 Upscale
Google's Omni v1.2 upscales a clip of up to 15 seconds to 720p, 1080p or 4K. The price depends on your clip's resolution and the target.
LTX 2.5 Alpha Gen
Lightricks LTX 2.5 Alpha Gen pulls the subject out of a video: a grayscale alpha matte plus a transparent ProRes 4444 cutout for your editor, at the clip's resolution up to 4K.
LTX 2.5 Reframe
Lightricks LTX 2.5 Reframe recomposes a video for a new aspect ratio (9:16, 1:1, 4:5, 5:4 or 16:9) and generates the footage the new frame needs, at 720p or 1080p, keeping the audio.
LTX 2.5 Upscale
Lightricks LTX 2.5 Upscale sharpens a video to 1080p, 4K or 8K and keeps its audio and frame rate.
LTX 2.5 SDR to HDR
Lightricks LTX 2.5 SDR to HDR turns standard footage into a 10-bit HLG HDR video at the clip's own resolution, with wider highlights, shadows and color.
Kling 1.6 Pro
Latest Kling 1.x with enhanced quality
MiniMax H3 Max Turbo Extend
MiniMax H3 Max Turbo for making videos longer, faster: upload a clip, describe what happens next, and it adds 5 to 15 seconds of new footage that continues the same scene. You pay only for the new seconds. Get the full extended video or only the new part, at 480p, 768p, 1080p or 2K.
Happy Horse 1.1
Upgraded cinematic short-video model — smoother motion, stronger instruction following, native audio behaviour, and multilingual lip-sync. Text-to-video, first-frame image-to-video, and reference-to-video (up to 9 reference images).
Grok Imagine Video 1.5 Lite
Grok Imagine Video 1.5 Lite — xAI's faster, lower-cost video model. Text-to-video and image-to-video clips of 1 to 15 seconds at 480p, 720p or 1080p.
VEED Dubbing
VEED Dubbing for video: translates the speech in a video of up to 10 minutes into one of 85+ languages, re-voices it in the speakers' own voices and syncs their lips to the new words.
Motion to Code GSAP
Motion to Code GSAP — turns a prompt, an image or a reference video of up to 15 seconds into an editable GSAP animation: one self-contained HTML page you can play, tweak and ship, plus a video preview. Flat motion design and kinetic typography work best.
Vidu 2
High-quality video generation with strong motion
Volcengine Lip Sync
Video-to-video lip sync — aligns mouth movements in an existing video with target audio. Multilingual dubbing supported.
Bria Video BG Removal (Green-Screen Despill)
Bria's enterprise-grade licensed video background removal with green-screen despill. Remove backgrounds from video clips cleanly, with green-screen spill correction for professional keying results.
Pixelcut Looping Video
Pixelcut Looping Video for product photos: turns one image into a seamless 5 to 15 second loop that starts and ends on it, with subtle motion or a full 360° spin, at 480p, 768p or 1080p.
Happy Horse
Cinematic short-video model — superior motion, film-grade lighting, strong prompt adherence, and multi-shot storytelling. Single text-to-video + image-to-video endpoint.
Video Upscaler
Enlarge and sharpen a video to a higher resolution. Choose a scale factor from 1x to 8x (default 2x).
Kandinsky 6.0 VSR
Kandinsky 6.0 VSR — Kandinsky Lab's video super-resolution model. Upscales generated or real footage 2x, 2.25x or 4x with sharper detail and stable, flicker-free frames (the first 5 seconds of a clip).
Motion to Code GLSL
Motion to Code GLSL — turns a prompt, an image or a reference video of up to 15 seconds into a seamlessly looping GLSL fragment shader, ready for Shadertoy, with a page that plays it live and a video preview. Best for abstract, generative and atmospheric loops.
Pika 2
Pika's latest video generation model
MiniMax Video 01
Fast video generation from images
Video Background Removal
Remove the background from a video and get a clean subject cutout, ready to composite onto any scene. Optional audio passthrough and background color.
LatentSync
Latent-space lip sync model
Seedance 2.0 Mini
Fast, budget-friendly video model — text-to-video and image-to-video (first frame, last frame, or reference) with optional synced audio. Great for quick drafts and high-volume social clips.
Video Translator
Translate short videos — transcribes the speech and generates dubbed audio plus subtitles in your target language. Up to 6 minutes per clip.
Music Video
Render a shareable video for a track, with optional artist and site credits.
Kandinsky 6.0 Lite
Kandinsky 6.0 Lite — Kandinsky Lab's fast, low-cost video model for quick iteration, with synchronized audio. 5-second text-to-video and image-to-video at 480p, or upscaled to 1080p or 4K.
Seedance 2.0 Fast
Faster, cheaper variant of Seedance 2.0 — text-to-video and image-to-video.
Kling 1.0 Standard
Standard-tier video generation model
Veo 3.1 Lite
Veo 3.1 Lite — budget-friendly video generation with text-to-video, image-to-video, and reference-to-video modes
Hailuo 2.3 Pro
MiniMax Hailuo 2.3 Pro high-quality video generation
Hailuo 2.3 Standard
MiniMax Hailuo 2.3 Standard video generation
Hailuo 02 Pro
MiniMax Hailuo 02 Pro video generation
Wan 2.6 Flash
Wan 2.6 Flash fast image-to-video
Topaz Video Upscale
Topaz AI video upscaling
Audio Models
22 models available
ElevenLabs Eleven v4
ElevenLabs' most expressive text-to-speech. Direct the delivery with inline audio tags like [whispers] and [laughs], shift emotion mid-sentence, and speak in 100 languages.
ElevenLabs Multilingual V2
High-quality multilingual text-to-speech
OpenAI TTS HD
Premium text-to-speech model
HeyGen Voice
HeyGen Voice — HeyGen's text-to-speech. Natural speech in 39 languages from 24 library voices, or clone a voice instantly from a short recording of one speaker.
Suno Music
AI music generation — create songs, instrumentals, and soundtracks
ElevenLabs Eleven v4 Turbo
The low-latency Eleven v4 from ElevenLabs. The same voices and inline audio tags as Eleven v4, at half the price per character.
Suno Speech
Suno Speech — Suno's expressive speech model for voice-overs, podcasts, audiobooks and ads. Describe a scene and Suno writes and performs the script, or bring your own script of up to 5,000 characters with tone direction, a female or male voice, variety and an optional original music bed. Two takes per request.
ElevenLabs Turbo V2.5
Fast text-to-speech with good quality
Suno Cover
Remix a song in the same style: upload a track or link one, describe the changes, and Suno returns two new takes.
HeyGen Voice Turbo
HeyGen Voice Turbo — the fast, low-cost tier of HeyGen Voice. The same 24 library voices, 39 languages and instant voice cloning from a short recording.
OpenAI TTS
Text-to-speech model
F5-TTS
High quality text-to-speech
Google Lyria 3.5
Google DeepMind's latest music generation model. Generate almost any type of music from text descriptions, with optional image inspiration.
VEED Clean Audio
VEED Clean Audio for recordings: removes background noise, hum and room echo from up to 30 minutes of speech and evens out the loudness, returning FLAC or WAV.
Gemini 3.8 Flash TTS
Google's Gemini 3.8 Flash text-to-speech. 30 voice presets, inline vocal cues like <laugh> and <sigh>, and a separate delivery-style prompt.
Gemini 3.1 Flash TTS
Gemini 3.1 Flash text-to-speech. 30 voice presets, 80+ languages, inline expressive tags ([sigh], [laughing]), style instructions up to 4K chars, and multi-speaker support.
Vocal Separation
Split a track into vocal and instrumental stems, or into up to twelve individual instrument stems.
Extend Music
Continue an existing track from any point, keeping its style.
Add Vocals
Sing over an instrumental you upload.
Add Instrumental
Build a backing track under vocals you upload.
Replace Section
Re-generate one time range of a track, leaving the rest untouched.
Gemini 3.8 Flash Lite TTS
The lower-cost Gemini 3.8 text-to-speech model from Google. The same 30 voice presets, vocal cues and delivery-style prompt as Flash.
3D Models
7 models available
Meshy 7.1 Multi-Image
Reconstruct a 3D mesh from up to four views of the same object. Resolves the sides a single photo can't see, with optional 2K geometry. Downloads as GLB.
Meshy 7.1
Turn a single photo into a textured 3D mesh, now with 4K geometry for the finest surface detail. Standard, low-poly or smart-topology meshes, optional PBR maps and auto-rigging. Downloads as GLB.
Meshy V7 Multi-Image
Reconstruct a 3D mesh from up to four views of the same object. Resolves the sides a single photo can't see. Exports GLB, FBX, OBJ, USDZ and STL.
Meshy 7.1 Text-to-3D
Generate a textured 3D mesh from a text description, now with 4K geometry for the finest surface detail. Controllable topology and polygon count, optional PBR maps and auto-rigging. Downloads as GLB.
Meshy V7
Turn a single photo into a textured 3D mesh. Controllable topology and polygon count, optional PBR maps and auto-rigging. Exports GLB, FBX, OBJ, USDZ and STL.
Meshy V7 Text-to-3D
Generate a textured 3D mesh from a text description. Controllable topology and polygon count, optional PBR maps and auto-rigging. Exports GLB, FBX, OBJ, USDZ and STL.
Tripo P2 Text-to-3D
Tripo P2 turns a text description into a textured 3D model with PBR materials, as a GLB. Choose quad or triangle topology and a face budget.