header-langage
简体中文
繁體中文
English
Tiếng Việt
한국어
日本語
ภาษาไทย
Türkçe
Scan to Download the APP
Partnered With:

Ideogram 4.6 focuses on precise image editing: continuous modifications without deviation, and native 4K image output

Beating AI News Flash: AI image generation company Ideogram releases the 4.6 model, supporting native 4K image generation without the need for post-upscaling. Compared to the previous generation's native 2K, details such as pores and fine lines on human skin, as well as textures of fabric and metal, are richer.


The new version continues to strengthen the precise editing capabilities that were the highlight of Ideogram 4.5. For example, if you only want to change the color of a piece of clothing, the AI will try not to touch the person's face or the background. Even after multiple consecutive edits, it can reduce color, texture, and detail shifts in other areas of the image.


According to official tests under the same conditions, the model's usability rate for changing objects to a specified color code increased from 46% in the previous generation to 83%; the usability rate for repairing product photo details increased from 61% to 86%; and the usability rate for modifying only a specified area increased from 82% to 90%.


On the Rapidata multi-turn image editing leaderboard, Ideogram 4.6 ranks first with 1,892 points, ahead of ByteDance's Seedream 5.5 and FLUX 3 Image. The evaluation had the model perform 5 consecutive edits on 199 real-person photos, after which humans judged whether the person's face remained unchanged.


The model is currently available on the Ideogram app, API, and multiple partner platforms. At the highest quality setting on the API, native 4K generation costs $0.20 per image, and 4K editing costs $0.60 per image.


Ideogram also stated that it will skip 4.5 and directly open-source the model weights of 4.6.

举报 Correction/Report
RelatedRecommended
According to BlockBeats' statistics, there were 13 financing deals last week with a total amount of approximately 376 million US dollars and an average financing amount of 28.9385 million US dollars.
Correction/Report
Submit
Add Library
Visible to myself only
Public
Save
Choose Library
Add Library
Cancel
Finish