Every model, one canvas

OpenAI's fast image model with Image 2-level quality, precise editing, and improved subject preservation

OpenAI's quality-focused image model for detailed generation, precise editing, and subject preservation

MiniMax's H3 Max via fal.ai: text-to-video, image-to-video, and reference-to-video with native audio at 480P or 768P.

OpenAI's GPT Image 2 with native reasoning, up to 4K output, and multi-image consistency across a batch.

OpenAI's detailed image model for precise edits and preserving products, characters, and brand details

Google's Nano Banana 2.1 (Gemini Nano Banana 2.1): the successor to Nano Banana 2 for text-to-image and image editing at 1K, 2K and 4K, with ratios from 1:8 to 8:1 (supports up to 14 input images).

Fast Gemini 2.5 Flash image variant for text-to-image generation and image mixing (supports up to 3 input images).

Kuaishou's Kling 4.0: text-to-video, image-to-video with an optional end frame, reference-to-video from up to 10 images, up to 7 characters (elements) and up to 3 voices, and prompt-driven video editing, all with native audio, 3-30s at 720p or 1080p.

Kuaishou's faster, lower-cost Kling O3 tier for text-to-video and image-to-video generation with native audio and clips up to 15 seconds.

PixVerse V6 for low-cost text-to-video and image-to-video, silent or with native audio, 1–15 second clips, and output up to 1080p.

Alibaba's WAN 3.0: top-ranked text-to-video and image-to-video with native audio and an optional end frame, 2–30s clips at 480p, 720p or 1080p.

Preview of Gemini 3 Pro image generation for text-to-image and image mixing (supports up to 14 input images).

OpenAI's GPT Image 1.5 model for image generation and edits (supports up to 10 input images).

Recraft's V4 text-to-vector model. Generates true, editable SVG graphics — logos, icons, and brand assets that stay sharp at any resolution.

MiniMax's H3 Image: detailed images that follow the prompt closely, at any size up to 20 megapixels, and edits that combine up to 16 reference images.

OpenAI's fast image model for interactive generation, improved layouts, and consistent edits

HeyGen's video model, built on MiniMax H3: text-to-video, image-to-video and reference-to-video with its own sound, 5–15s clips at 480p, 768p, 1080p or 2K.

Ideogram's Vector 1.0. Generates editable SVG logos, icons, lettering and illustrations from a prompt, or converts one input image into SVG (the prompt is not used then).

Google's Gemini Omni 1.2 Flash: text-to-video, image-to-video with an optional last frame, and video from up to 6 character, product or style reference images, with native audio and lip-synced speech, 3–15s clips from 360p to 4K.

MiniMax's H3 Max Turbo via fal.ai: the same H3 Max schema at half the per-second price, optimized for latency. No reference-to-video endpoint.

Bria's commercially licensed product-compositing model for placing product cutouts into scenes with controlled position, perspective, and natural lighting.

Bria's Fibo 2.0: text-to-image and targeted edits of up to 4 images (background swaps, product restaging, lighting changes), trained only on licensed data.

Professional FLUX 1.1 model with enhanced quality and capabilities.

Black Forest Labs' FLUX 3 Image: detailed, well-composed images with precise layout and text, up to 4K, and precise edits combining up to 10 reference images.

fal's post-trained FLUX 3: text- and image-to-video with native audio, an optional end frame, stronger prompt adherence, and 5–20s clips at 720p or 1080p.

Advanced FLUX model for image generation and editing with reference image support for context and composition guidance.

Developer-focused FLUX.2 variant with lower latency and go_fast toggle.

Flexible FLUX.2 variant optimized for creative exploration with tunable steps and guidance.

Un-distilled FLUX.2 Klein 4B base model optimized for fine-tuning and multi-reference workflows.

Un-distilled FLUX.2 Klein foundation model for flexible text-to-image and multi-reference workflows.

Nano Banana 2 Lite — Google's fastest, most cost-efficient Nano Banana image model. Near-real-time 1K text-to-image and image editing (supports up to 14 input images).

xAI's Grok Imagine Image 2.0: text-to-image and edits of up to 5 images, at low or medium quality and 1k or 2k resolution.

xAI Grok Imagine 1.5 image-to-video: animate a source image with a text prompt at 480p or 720p.

The highest quality Ideogram v3 model. v3 creates images with stunning realism, creative designs, and consistent styles

Turbo is the fastest and cheapest Ideogram v3. v3 creates images with stunning realism, creative designs, and consistent styles

Ideogram's latest text-to-image model. Best-in-class text rendering for posters, logos, and signage, with fine detail and strong creative control.

Google's flagship Imagen 4 model for high-quality image generation with improved text rendering

Kuaishou's Kling v1 text-to-video with pre-baked cinematic camera templates (dolly/crane, orbit, pan, tilt, roll, zoom) via fal.ai. Camera control is a Kling v1-era feature; newer Kling tiers omit it.

Kuaishou's Kling v1.5 Pro image-to-video with motion-brush trajectory pathing — paint per-region motion paths over a start image (dynamic masks + a static hold region) via fal.ai.

Krea 3: photoreal text-to-image and edits combining up to 8 reference images. quality picks how much Krea reasons (high and ultra plan edits more carefully); turbo_mode trades quality for speed.

Decomposes a single image into multiple editable RGBA layers (foreground, background, text, and individual elements) in one pass, so each piece can be moved and edited independently on the canvas.

The faster variant of MiniMax's H3 Image, at about twice the price: images at any size up to 20 megapixels, and edits that combine up to 16 reference images.

High-quality text-to-image model from Qwen with support for multiple canvas dimensions and LoRA weights.

Qwen's enhanced image editing model supporting multi-image conditioning and rich prompt controls.

Recraft's V4 text-to-image model. Balanced composition and cohesive color for design and marketing assets; top-ranked on the Text-to-Image Arena.

ByteDance's Seedance 2.5 via fal.ai: single-take clips up to 30 seconds at 480p, 720p, or 1080p with joint audio, plus image-to-video and reference-to-video.

ByteDance's SeedDream 4 model for high-quality text-to-image and image-to-image generation with support for up to 4K resolution.

ByteDance's SeedDream 4.5 model for high-quality text-to-image and image-to-image generation with improved spatial understanding and world knowledge, supporting up to 4K resolution.

ByteDance's flagship Seedream 5.0 Pro for structured image generation and region-precise editing, with multilingual typography and up to 10 reference images.

Preview release of Veo 3.1 supporting enhanced text-to-video and image-to-video generation on Vertex AI.

Veo 3.1 Fast Preview delivers rapid preview renders for text-to-video and image-to-video via Vertex AI.