GPT Image 2

State-of-the-art image generation and editing. Exceptional at instructions, text rendering and identity consistency.

The pro tier: sharper detail, up to 4K, multi-image reference (up to 8) and complex multi-panel compositions with reliable text.

ByteDance's multimodal image model for controlled generation, precision editing, realistic portraits and multilingual text. Text-to-image, or drop an image to edit — up to 2K.

ByteDance's refined image model for 4K generation, precise editing and consistent multi-image output. Text-to-image, or drop an image to edit — up to 4K.

ByteDance's next-gen model combining text-to-image, image-to-image and editing with batch consistency, high speed and professional output. Drop up to 10 images to edit — up to 4K.

Black Forest Labs' Flux 2 for photoreal detail, strong multi-reference consistency and accurate text rendering. Text-to-image, or drop up to 8 reference images — 1K ($0.025) or 2K ($0.035).

xAI's Grok Imagine for image generation with strong visual consistency. Text-to-image returns a batch per run (Standard: 6 images $0.02 · Quality: 4 images $0.025), or drop up to 5 images to edit ($0.02/image).

OpenAI's flagship image model for high-quality creation and precise editing, with strong instruction following and improved text rendering. Text-to-image, or drop up to 16 reference images — Medium ($0.02) or High ($0.11).

Restore and upscale images to 4K — reconstructs real detail instead of blurring pixels.

Outpainting-based resizing: change aspect ratio while the model naturally extends the scene.

Instant background removal with hair-level edge accuracy. Transparent PNG output.