Image generation

Ideogram 4.0

by Ideogram

Ideogram's open-weight image model. Frontier in-image text, layout control, and 2K output.

Ideogram 4.0

Key features

Technical specifications

Open

Open weights under a commercial license

0.97 OCR

X-Omni English OCR for in-image text

16 colors

Condition output on up to 16 hex colors

Up to 2K

256 to 2048 px per side, flexible ratios

Use cases

Posters and packaging

Design where the title, tagline, and small print all have to read correctly. Text renders legibly, not as shapes.

Multilingual campaigns

Localize one visual across markets by swapping the text per language while layout and palette stay fixed.

Brand-locked visuals

Feed the brand's hex palette into the prompt and every generation stays on-brand, tile to banner.

Unusual formats

One set of weights covers square thumbnails, widescreen, 2048 by 768 ultrawide banners, and social headers.

Programmatic generation

JSON prompts are built for code. Generate catalogs or ad variants from a script, each element typed and validated.

Self-hosted pipelines

Teams that can't use a third-party API can fine-tune the open weights and run them in their own infrastructure.

Prompt examples

Event poster

Event poster

Jazz festival poster, bold title at top, lineup text readable at the bottom

Edit prompt
Packaging

Packaging

Coffee bag front panel, roastery name in serif type, warm morning light

Edit prompt
Brand palette

Brand palette

Product launch banner locked to a teal, sand, and rust brand palette

Edit prompt
Ultrawide banner

Ultrawide banner

Ultrawide website banner, mountain ridge at dawn, headline on the left

Edit prompt
Multilingual sign

Multilingual sign

Tokyo storefront with accurate Japanese signage, soft rain, evening glow

Edit prompt
Magazine cover

Magazine cover

Architecture magazine cover, masthead at top, coverlines down the right

Edit prompt

Simple pricing

Get started for free today, with the option to upgrade or cancel anytime.

Basic

$0/ month
billed as $0 per year

900 monthly credits

1 user only

All models

Workflows

Standard

$0/ month
billed as $0 per year

3200 monthly credits

1 user only

All models

Workflows

Pro

$0/ month
billed as $0 per year

6200 shared monthly credits

1 user

+ up to 4 more at extra cost

All models

Workflows

Pro Max

$0/ month
billed as $0 per year

24000 shared monthly credits

1 user

+ up to 9 more at extra cost

All models

Workflows

Enterprise

For higher limits

Custom

pricing and billing terms

Unlimited credits
Custom seat limits
All models
Workflows
Pricing Gradient

Free

For playing around

$0

forever free

Up to 20 credits
1 user only
Limited models
Workflows
Compare plan details

FAQs

What is Ideogram 4.0?
Ideogram 4.0 is a 9.3 billion parameter open-weight text-to-image model from Ideogram, released on June 3, 2026. It focuses on accurate in-image text rendering, bounding-box layout control, color palette conditioning, and output up to 2K, with weights available to download under a commercial license.
Is Ideogram 4.0 open source?
Ideogram 4.0 is open-weight rather than fully open source. The weights, inference code, and prompting guide are public on Hugging Face and GitHub, and commercial deployments are covered by a license that scales with usage. You can download, fine-tune, and self-host the model.
How good is Ideogram 4.0 at rendering text?
Text rendering is the model's headline strength. Ideogram reports a 0.97 score on the X-Omni English OCR benchmark, which measures whether text inside a generated image is actually readable and correctly spelled, and the model handles multilingual text as well as English.
How does layout control work in Ideogram 4.0?
You attach bounding boxes to elements in the prompt, each coupled to a plain-language description, and the model places those objects inside their requested regions. Ideogram reports a 0.69 mIoU score on the 7Bench layout benchmark, which measures how tightly generated objects sit inside their boxes.
What is structured JSON prompting?
Instead of one long sentence, an Ideogram 4.0 prompt is a JSON object: a high-level scene description, a style block for aesthetics and lighting, individual elements with optional bounding boxes, typed text elements with the literal string to render, and an optional palette of up to 16 hex colors. The reference pipeline validates each prompt against the schema before generating.
What resolutions does Ideogram 4.0 support?
Ideogram 4.0 generates from 256 to 2048 pixels per side with flexible aspect ratios. A single set of weights covers 1024 square output, 1536 by 1024 landscape and portrait, 1920 by 1088 widescreen, 2048 by 768 ultrawide, phone wallpapers, and 1584 by 396 social banners.
What are the Turbo, Default, and Quality presets?
They are three sampling presets that trade speed for polish: Turbo runs 12 denoising steps, Default runs 20, and Quality runs 48. A common workflow is drafting compositions on Turbo, then re-running the chosen prompt on Quality for the final asset.
Can I run Ideogram 4.0 on my own hardware?
Yes. Ideogram publishes fp8 and nf4 quantized checkpoints that fit on a single 24 GB GPU, alongside the full weights and inference code. Teams can also fine-tune the model on their own brand or product data and deploy it inside their own environment.
How does Ideogram 4.0 compare to GPT Image 2?
On Ideogram's designer-preference ELO leaderboard, Ideogram 4.0 scores 1062, second behind the closed-source GPT Image 2 at 1141 and ahead of every other open-weight model. The practical difference is access: GPT Image 2 is API-only, while Ideogram 4.0 can be downloaded, fine-tuned, and self-hosted.

Other models

Reve 2.0

Reve AI

Reve AI's layout-first image model. Place every element by hand, edit the result like a design file, and render crisp text at up to 4K.

Bernini

ByteDance

ByteDance's open-source video model for instruction-based editing, with the rest of the frame locked and subject identity held.

Grok Imagine v1.5

xAI

xAI's image-to-video model with native synchronized audio. Animate any still into a clip with sound, dialogue, and music.

Veo 4

Google DeepMind

Google DeepMind's next video model. Native 4K, longer clips, multi-shot character consistency, and a cinematic camera language in a single prompt.