ByteDance Releases Doubao Image Creation Model Seedream 5.5, Integrated into Doubao, Doubao Work, and Jimeng

On November 9, ByteDance officially released the Doubao image creation model Seedream 5.5. Compared with the previous generation, Seedream 5.0 Pro, the new model offers significantly faster generation speeds, while also enhancing multi-turn editing stability, natively supporting transparent asset generation, and focusing on optimizing image realism and small-text rendering, better meeting the quality, stability, and efficiency requirements of professional creation scenarios.

Currently, Doubao, Doubao Work, and Jimeng have all integrated Seedream 5.5, and the API will soon launch on Volcano Ark. In terms of pricing, the Seedream 5.5 API continues the two-part billing structure of "input image + output image," with no charges for text tokens. Taking 2K square image editing as an example, the cost per image is around 0.28 yuan, less than half that of Seedream 5.0 Pro.

In multi-turn image editing, Seedream 5.5 has stronger instruction-following capabilities and achieves pixel-level preservation of non-edited areas. In actual tests, with the model using the result of each round as the input for the next, after completing hundreds of consecutive rounds of modifications, the subject features, composition proportions, and lighting style outside the specified editing areas remained consistent.

After 100 rounds of modification, pixel-level preservation is achieved in non-edited areas

Seedream 5.5 natively supports the generation and separation of RGBA transparent assets, can directly output images with an Alpha channel, and supports layered asset output. In addition to text-to-image, creators can also generate the required transparent assets from reference images; while preserving the rich forms and textures of the subject, the model accurately handles contour edges and transparent areas, allowing them to be reused across scenarios as design assets and improving production efficiency.

Seedream 5.5 natively supports transparent asset generation

In addition, Seedream 5.5 improves the representation of lighting, materials, and skin texture in realistic scenes, reducing excessive glossiness and overexposure, so that generated results are closer to natural photography. In scenarios such as backlit portraits, visual effects compositing, and multi-image fusion, the spatial relationships, occlusion layers, and lighting transitions between newly added elements and the original image are more reasonable.

Generated results are closer to natural photography

Seedream 5.5's small-text rendering has also improved, especially for Chinese text generation at smaller font sizes, where the structure of Chinese characters with complex strokes remains clearly legible, making it usable for content with high text precision requirements such as infographics and knowledge diagrams.

Seedream 5.5 is being fully integrated into professional-grade visual creation workflows. Based on specialized test results from the MagicArena platform, Seedream 5.5's usability rate in six scenarios—e-commerce, marketing, film and television, gaming, office, and fashion design—showed significant improvement over Seedream 5.0 Pro.

Image source: ByteDance Seed official website

In e-commerce marketing, users can leverage the model's multi-turn editing capabilities to produce multiple outfit effects for the same model at lower cost; in home decoration design, designers can quickly validate different decoration schemes while keeping the spatial layout and details unchanged; in film and television storyboarding and game asset production, creators can also continuously adjust a single image without losing image quality.

ByteDance stated that a Pro version of the Seedream 5.5 series will be released later, which will support more complex context inputs and creation instructions in the future, continuously improving the model's performance in professional tasks. At the same time, the team will expand the model's application in Agent workflows around the actual production needs of fields such as film and television, e-commerce, advertising, and design, and continue to explore more general visual intelligence.

Leiphone Leiphone