Ideogram 4.0 drops as an open-weight model with native 2K resolution and improved text rendering
Quick Answer
Ideogram 4.0 has been released as an open-weight text-to-image model featuring native 2K resolution and enhanced text rendering.
Quick Take
It ranks first among open models on the DesignArena leaderboard, surpassing all but closed systems from OpenAI and Google. A paid license is required for commercial use.
Key Points
- Ideogram 4.0 features native 2K resolution and improved text rendering capabilities.
- Ranks first among open models on the DesignArena leaderboard.
- Only closed systems from OpenAI and Google perform better.
- Commercial use of Ideogram 4.0 requires a paid license.
- Bounding box control is a new feature in this version.
📖 Reader Mode
~2 min readIdeogram has released version 4.0 of its text-to-image model as an open-weight model.
According to Ideogram, the new features include native 2K resolution, transparent backgrounds, precise layout control via bounding boxes, and improved text rendering in images, useful for logos and posters. Editable text and layers are coming soon, the company says.
The model can run on your own hardware and be fine-tuned with your own data. Weights and code are available for download on GitHub, but commercial use requires a paid license.
According to the DesignArena leaderboard, Ideogram 4.0 ranks first among all open-weight models. Only closed models from OpenAI and Google score higher. In the text-to-image arena, it also takes first place in quality mode and ninth overall. The model is available in three quality tiers via Ideogram's own hosted API, according to the Ideogram website:
| Quality level | Price per image |
|---|---|
| Turbo | 0.03 dollar |
| Default | 0.06 dollar |
| Quality | 0.10 dollar |
Ideogram 4.0 is also available on the web and across partner platforms, including Hugging Face, ComfyUI, fal, Runware, Magnific, Krea AI, Leonardo AI, Picsart, Cloudflare, Replicate, Gamma, Flora AI, and Kittl. In our benchmark prompt, the model easily outperforms Midjourney v8, lands roughly on par with Flux, but falls short of GPT-Image-2, Nano Banana Pro, or Luma Uni-1.1. That's just one prompt, though, and it mainly tests prompt following and the model's ability to render abstract concepts unlikely to appear in the training data, like a horse-riding astronaut. As always, your own testing is a must.
— Originally published at the-decoder.com
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from The Decoder
See more →
An AI model programmed nonstop for 19 days on a single MirrorCode task that cost $2,600 to run
Epoch AI's MirrorCode benchmark reveals Claude Opus 4.7 as the leader with a 56% solve rate, reconstructing a 16,000-line toolkit in 14 hours. Despite this, all models tested struggle with the most complex tasks, highlighting limitations in current AI capabilities. The single task consumed $2,600 over 19 days, raising questions about cost-effectiveness in AI development.
