Ideogram 4

by Ideogram

Text-to-image model by Ideogram. 9.3B params, best-in-class typography and text rendering among open weights. Single-stream DiT paired with a Qwen3-VL-8B text encoder, loaded as conditional + unconditional models. Day-0 ComfyUI support. Non-commercial license.

ImageNon-commercialReleased Jun 20269.3B parameters

Hugging Face repo ↗Ideogram ↗

Drag or click to move along the pipeline

full qualitylightest

Lighter tiers fit smaller GPUs and lose some detail. Start with the heaviest one that loads.

The model

Ideogram 4

FP8from 18.8 GB

Three pieces, three separate downloads.

Diffusion modelGoes in diffusion_models/

Generates the image

Pick one: smaller loads faster, larger keeps more detail.

8.6 GB↓↗

Text encoderGoes in text_encoders/

Turns your prompt into guidance

9.9 GB↓↗

VAEGoes in vae/

Decodes the result into pixels

0.3 GB↓↗

LoRAs (optional)Goes in loras/

No LoRAs indexed for this model.

SamplerKSampler

Copy these values into your KSampler.

Steps20

Samplereuler

SchedulerIdeogram4Scheduler

Native res.1024x1024 (native), up to 2048x2048

Guidance7

Usage notes

GGUF is the lightest way to run this model.

Non-commercial. Check the official repo for usage terms.

Click a file name to copy it.