ideogram-4-nf4-diffusers

Ideogram 4 is Ideogram's first open weight text-to-image model. It is a state-of-the-art foundation model trained from scratch — not a fine-tune of any existing model. 

Key Features:

• Prompting: While plain-text prompts work, the model performs best with structured JSON captions. A "magic prompt" LLM can expand plain text into structured JSON.
• Text Rendering: Best-in-class multilingual text rendering with high fidelity for signage, logos, captions, watermarks, and multi-line text directly from the prompt.
• Layout Control: Supports explicit bounding-box coordinates in prompts to place subjects, text elements, and background regions.
• Color Palette: Users can specify hex colors in the prompt to control the image\'s dominant color scheme.
• Resolution: Native support for resolutions from 256 to 2048 pixels (multiples of 16) with aspect ratios up to 6:1. The noise schedule auto-adjusts per resolution.

---

The model is a component of the image generation pipeline, consisting of:

• Text encoder: ~4.5B parameters,
• Transformer: ~9.6B parameters,
• Unconditional transformer: ~4.8B parameters,
• VAE: ~84M parameters, 

Total: ~18.9B parameters

Announce Date: 03.06.2026
Parameters: 10B
Developer: Ideogram
Diffusers Version: 0.39.0.dev0
License: ideogram-4-non-commercial

Use our pre-built public endpoints for free to test inference and explore ideogram-4-nf4-diffusers capabilities. You can obtain an API access token on the token management page after registration and verification.

Rent your own physically dedicated instance with hourly or long-term monthly billing.

We recommend deploying private instances in the following scenarios:

• maximize endpoint performance,
• enable full context for long sequences,
• ensure top-tier security for data processing in an isolated, dedicated environment,
• use custom weights, such as fine-tuned models or LoRA adapters.

Contact our dedicated neural networks support team at nn@immers.cloud or send your request to the sales department at sale@immers.cloud.