Krea

Krea 2 Large

More powerful version of Krea 2 optimized for expressive photorealism.

Krea 2 Medium

A smaller variant of Krea 2. Works best with illustrations and graphic design

Krea 2 Turbo

The fastest Krea 2 model. Best for quickly iterating on expressive illustrations.

ChatGPT 2

Smartest model. Excellent prompt following. Best for complex diagrams, characters, and image editing.

Nano Banana Pro

Smartest model. World's best prompt adherence. Best model for complex tasks and image editing.

~100

Nano Banana 2

Google's latest flash image model (also known as Gemini 3.1 Flash Image) optimized for fast generation with support for up to 4K resolution.

Nano Banana 2 Lite

Fast access to Google Nano Banana 2 Lite for text-to-image generation.

Luma UNI-1

Luma UNI-1 reasons through prompts before generating and supports optional visual references.

Ideogram 4.0

Open image model focused on design, text rendering, and 2K photorealistic output.

Seedream 5 Lite

Latest high-quality model from ByteDance with deep thinking, web search capabilities, and up to 14 style images.

~25

Seedream 5 Pro

Flagship ByteDance model with precise local edits, multi-image fusion, strong text rendering, and up to 10 reference images.

~35

Ideogram 4.5

Generate images and make edits with up to five image references.

Muse Image

Muse Image plans before it renders: it searches the web for real references, runs code for precise elements like charts and QR codes, and handles text-to-image, editing, and multi-reference composition (up to 10 images) in one model.

Grok Imagine 2.0

Grok Imagine 2.0 generates images from text and edits up to 3 reference images. Choose a quality level and 1K or 2K output to trade detail against cost.

Recraft V4

Sharp, detailed images from Recraft with Standard and Pro modes.

Recraft V4.1 Flash

Recraft’s fast tier for raster art, illustration, and photography.

Qwen Image 3

Alibaba’s third-generation image model. Reads prompts up to 4.5k tokens and renders text down to 10px across 12 languages, which suits posters, infographics, and storyboards. Attach up to 3 reference images to edit instead of generate.

Qwen Image 2.1

Generates and edits in one model, with native transparency for stickers, cutouts, and layered art. Keeps faces and products recognizable through outfit, material, and background changes, and composes from multiple references.

Free New

Z Image

Cheapest model. Medium quality photorealism at a budget. Realistic textures, weak diversity.

Free New

Flux 2

FLUX.2 [dev] from Black Forest Labs. Enhanced realism and crisper text generation.

Krea 1

Most creative model with LoRAs. Excels at artistic photorealism. Realistic textures, weaker structure.

GPT Image 2.5 Flare

GPT Image 2.5 Flare generates images with transparent background support, following the gpt-image-2 request shape.

Nano Banana

Smart model optimized for precise prompt adherence. Best for most tasks.

Flux 2 Klein

A lightweight 4 billion parameter version of Flux 2, optimized for speed while maintaining quality. Supports up to 2K resolution.

Luma UNI-1 Max

Luma UNI-1 Max reasons more deeply than UNI-1 for higher-quality results and supports optional visual references.

Qwen Image 2512

Latest Qwen model with enhanced human realism, finer natural detail for landscapes and textures, and improved text rendering with better layout and composition.

ChatGPT 1.5

Smartest model. Excellent prompt following. Best for complex diagrams, characters, and image editing.

GPT Image 2.5 Sunburst

GPT Image 2.5 Sunburst is the faster GPT Image 2.5 variant. Five quality tiers trade latency against fidelity.

Seedream 4

Latest high-quality model from ByteDance with flexible resolution support

Flux.1 Dev

Best model for LoRA's. Fastest model. Highly customizable less realistic colors and textures.

Seedream 4.5

Quality 4K model optimized for vibrant photorealism.

Flux 2 Pro

Medium quality model by Black Forest with stable visuals.

~55

Flux 2 Flex

BFL's next generation model, excelling at rendering text and fine details.

Flux 2 Max

Frontier model by Black Forest Labs. Most capable Flux 2 model.

~350

ChatGPT Image

Older ChatGPT image model. Intelligent prompt adherence with sharp semi-realistic output.

Wan 2.2 Image

Cinematic outputs with crisp textures and realistic color palette, but weak diversity and structure,

Qwen Image

Semi-realistic model with great text generation and prompt adherence. Capable of following long and highly detailed prompts.

Kling o1 Image

Kling Omni image model with support for up to 10 reference images with tags.

Flux Kontext Dev

Frontier model designed for image editing. Capable of advanced reasoning and style transfer.

Flux Kontext Pro

Frontier model designed for image editing. Capable of advanced reasoning and style transfer.

Runway Gen-4 Image

Cinematic image model with references. Best at combining multiple characters into a single image.

~35

Ideogram 3.0

Highly aesthetic, general-purpose model. Excels at flat and graphic styles.

~55

Flux 1.1 Pro

Advanced yet efficent model from Black Forest Labs.

Flux 1.1 Pro Ultra

Black Forest Lab's highest quality model text to image model.

Flux.1 Krea

Distilled and open sourced version of Krea 1. Better at text rendering and anatomy.

Try video models

Seedance 2.5

ByteDance's flagship video model: cinematic 30-second clips at up to 1080p with 20% better prompt adherence, native synchronized audio, tagged image/video/audio references, region-level editing, and frame animation.

~850

MiniMax H3

Minimax frontier video model rendering at 2K. Generate from text, animate between start and end frames, or condition on tagged image, video, and audio references.

~500

Gemini Omni Flash

Google's video-first multimodal model. Generate from text, a starting image, or tagged image references — with native speech and sound effects, plus single-clip video editing.

~750

Seedance 2.0

ByteDance frontier video model with cinematic motion, native synchronized audio, tagged image/video/audio references, and support for animating between start and end frames.

~1100

Kling 3.0

Latest frontier model from Kling with native audio and extended durations up to 15 seconds.

Veo 3.1 Lite

Faster and more affordable version of the powerful Veo 3.1 model.

Flux 3 Video

One model across image, video and audio. FLUX 3 renders up to 20 seconds in a single pass with multilingual speech, effects and ambience generated alongside the frames, and holds multiple scenes and camera angles together across cuts.

~650

MiniMax H3 Max

MiniMax H3 Max is a post-trained variant of H3 tuned for stronger prompt adherence and better aesthetics, served through fal. Text-to-video, image-to-video with optional end frame, or reference-to-video with tagged image, video, and audio references — up to 15 seconds at 480p, 768p, or native 1080p.

MiniMax H3 Max Turbo

MiniMax H3 Max Turbo is fal’s speed- and cost-optimized H3 Max variant. Text-to-video, or image-to-video with an optional end frame — up to 15 seconds at 480p, 768p, or native 1080p. No reference-to-video on this tier.

LTX-2.5 Pro

Lightricks LTX-2.5 in its quality-optimized mode, for final output. Generates synchronized native audio in the same pass, with start and end frame support and optional camera motion.

~750

LTX-2.5 Fast

Lightricks LTX-2.5 in its speed-optimized mode, built for rapid iteration. Generates synchronized native audio at every resolution, up to 20 seconds and 4K, with start and end frame support and optional camera motion.

LTX-2.3 22B

High-quality Lightricks video generation model with native audio and start and end frame support.

Kling 3.0 Turbo

Kling 3.0 Turbo text and image-to-video model with 720p/1080p output and durations up to 15 seconds.

Runway Gen-4.5

Latest frontier model from Runway with native text-to-video.

~450

Seedance 2.0 Fast

Faster ByteDance Seedance 2 variant with cinematic motion, native synchronized audio, tagged image/video/audio references, and support for animating between start and end frames.

~900

Seedance 2.0 Mini

ByteDance Seedance 2.0 Mini model with native audio, tagged image/video/audio references, and support for animating between start and end frames.

~550

Gemini Omni Flash 1.1

Google's latest Gemini Omni video model. Generate from text, a starting image, or tagged image references — with native speech and sound effects, plus single-clip video editing.

~750

Grok Imagine 1.5

Grok Imagine 1.5 creates clips with synchronized audio, music, and sound effects. Text prompts generate a K2 start frame before animation.

~600

HappyHorse

HappyHorse generates synchronized audio-video clips from text, image references, or a source video edit instruction.

~500

Flux Video Edit

Black Forest Labs' video editing model. Attach a clip up to 15 seconds and describe the change — remove, add or replace objects, restyle, change text, dialogue or setting. Length, framing and timing follow the source.

~100

Kling o3

Advanced reasoning video model (720p). Supports image, element, and video references for precise creative control.

Kling o3 Pro

Advanced reasoning video model (1080p). Supports image, element, and video references for precise creative control.

Wan 3.0

Alibaba's latest generation video model: clips up to 30 seconds at up to 1080p with native synchronized audio, start and end frames, and up to 10 reference images, 5 reference videos, and 5 reference audios.

~550

LTX-2 19B

Affordable medium-quality audio-video model from Lightricks with native sound generation.

~35

Vidu Q3

New model excelling at anime.

~550

Krea Realtime

Real-time video generation model. Instant results at very low cost and quality.

Hailuo 2.3

Frontier model with dynamic motion.

Hailuo 2.3 Fast

Cheapest medium-quality model. Best for most use cases.

Grok Imagine

Fast, high-quality video generation by xAI.

Seedance 1.5 Pro

Latest medium quality model from ByteDance with audio generation and end frame support.

Kling o1

Intelligent video model that thinks before generating. Supports image, element, and video references for precise creative control.

Sora 2

OpenAI's new powerful video model. Rich world knowledge with very stable structure.

Wan 2.6

Medium-quality model from Alibaba with improved quality and multi-shot support.

~350

Sora 2 Pro

OpenAI's most advanced intelligent video generation model.

~900

Seedance 1.0 Pro Fast

Fast and cheap model. Up to 12 seconds.

~85

Veo 3.1

Best video model. Highest-quality frontier model with audio and reference images.

~1200

Veo 3.1 Fast

Faster and more affordable version of the powerful Veo 3.1 model with audio.

~450

Wan 2.5

Latest medium quality model from Alibaba.

~350

Kling 2.6

Frontier model from Kling with native audio. Highest quality at a moderate price point.

Kling 2.5

Next-gen model with improved dynamics and enhanced style adaptation from Kling.

Kling 2.5 Turbo

Top-tier text-to-video generation with unparalleled motion fluidity and cinematic visuals.

Vidu Q2

High-quality model with reference images support.

Hailuo 02

Frontier model with dynamic motion.

Seedance 1.0 Pro

Fast, high-quality model from ByteDance.

Veo 3

Older version of the highest-quality frontier model Veo 3.1.

~1200

Veo 3 Fast

Older, faster and more affordable version of the leading Veo 3 model with audio.

~600

Ray 3.2

Luma's Agents API video model with optional start and end anchor frames.

Wan 2.1

Fastest low-quality model with video Lora support.

Wan 2.2

Fast, lower-quality model from Alibaba.

Seedance 1.0 Lite

Fast and affordable medium-quality model from ByteDance.

Kling 2.1

Frontier model with 1080p resolution.

Veo 2

Expensive high-quality model from Google.

~1900

Runway Gen-4 Video

Medium quality model with a focus on cinematic visuals but weaker structure and motion.

Runway Gen-3

Old generation cinematic model with higher consistency.

Kling 1.6

Previous generation high-quality model for complex scenes.

Kling 2.0

High-quality model with great aesthetics.

~1000

Hailuo

High-quality model with camera control.

MiniMax 01-Live

High-quality model for animating people.

Ray 2

Older medium model with natural motion from Luma Labs.