GIZMO / IMAGES

Every image model. One playground.

Compare 15 image models from Krea, OpenAI, ByteDance, Ideogram, Alibaba, Google, Black Forest Labs, Recraft side by side, edit with your own reference images, and save what you like to your library. The same models work through the API and inside your agents.

Open the image playground →Image API docs ↗

Image generation is part of Starter and above. Each image is billed at the provider’s price plus your plan’s fee, after your weekly included usage.

An intricate isometric cutaway of a three-story treehouse library with spiral stairs, tiny readers, lanterns and a rope bridge, crisp detailed illustrationGPT Image 2.5 SunburstA golden retriever wearing tiny round sunglasses, lounging on a pastel pool float in a turquoise pool, bright summer afternoon, playful photoNano Banana 2Product photo of a matte black mechanical watch on a slab of wet slate, water droplets, dramatic rim lighting, macro detail, luxury advertisingSeedream 5.0 ProMinimal logo sheet for a coffee brand called "Ember & Oak": an acorn with a flame, wordmark and two color variations on cream paperIdeogram 3A bustling night market in Taipei with glowing signs, steam rising from food stalls, people in motion, photorealistic, vibrant colorsQwen Image 3

The image models

Every sample below was made by the model itself, from the prompt on its page.

KreaKrea 2 MediumCreative illustration and photography. References guide visual style.ReferencesKreaKrea 2 LargeExpressive visual style and photorealism with optional style references.ReferencesOpenAIGPT Image 2.5 FlareFast image generation with detailed layouts and precise editing.EditsOpenAIGPT Image 2.5 SunburstPrecision-focused generation for intricate visual details.EditsByteDanceSeedream 5.0 ProMultimodal image generation and reference-guided editing.EditsIdeogramIdeogram 4Crisp graphic design and accurate text rendering.AlibabaQwen Image 3Multilingual typography, realistic scenes, and editing.EditsOpenAIGPT Image 2Detailed images, typography, and precise edits.EditsGoogleNano Banana 2Fast generation and reference-guided editing.EditsGoogleNano Banana ProHigh-fidelity compositions and text.EditsBlack Forest LabsFLUX.2 ProPhotorealism, product shots, and editing.EditsByteDanceSeedream 5.0 LiteHigh-resolution scenes and visual reasoning.EditsByteDanceSeedream 4.5Consistent reference images and detailed scenes.EditsIdeogramIdeogram 3Posters, logos, and typography. References guide style.ReferencesRecraftRecraft V4Graphic design, illustration, and brand visuals.

Three ways to make images

Playground

Write one prompt, pick several models, and compare the results side by side. Add reference images to edit or guide the style.

API

One request per image at /api/v1/images/generations. Poll for the result and receipt; images can be saved to your library.

Agents

Your agents can create images mid-conversation with the image model you choose, and reuse images from your library.

Claude & ChatGPT

Ask for it in a chat: connect Gizmo to Claude, ChatGPT or Codex over MCP. How to connect

Start creating →