Choose the right model
Choose HiDream-O1-Image, HiDream V1, or another HiDream model for your goal. Image models create images, while video models support text-to-video and image-to-video.
HiDream develops native omnimodal AI models for text, image, and video inputs, with image generation, video generation, audio-visual synchronization, physics-grounded understanding, and intelligent narrative planning.
Hover to reveal another image. Move away to restore. Tap on touchscreens.
Our most capable model yet. Sharper prompt fidelity, readable type in every layout, and edits that stay true to your intent from the first generation to the last revision.
HiDream is Zhixiang Future’s native omnimodal AI model system. It understands and processes text, images, and video, then transforms them into images, videos, and interactive virtual worlds. HiDream spans image generation, video generation, interactive world models, and embodied intelligence, with representative models including HiDream-O1-Image, HiDream-O1-World, HiDream-O1-Embodied, and HiDream V1. With multimodal intent understanding, physics-grounded modeling, audio-visual synchronization, and visual narrative capabilities, HiDream supports AI image creation, image-to-video, text-to-video, digital scene building, and intelligent content creation.
Choosing HiDream means working with a native omnimodal AI model system that spans images, video, interactive worlds, and embodied intelligence. HiDream understands text, images, and video, then follows user intent from content planning through generation. Compared with models focused only on visual output, HiDream provides broader capabilities for complex instruction following, subject consistency, text rendering, physical behavior, shot planning, and audio-visual synchronization. HiDream V1 can also plan video duration around events, action cycles, and narrative rhythm, generating 5–20 second 1080p videos for AI images, image-to-video, text-to-video, digital scene building, and visual content creation.

Connect objects, settings, and context with the details of your imagination.
Creating with HiDream usually takes only a few steps.
Choose HiDream-O1-Image, HiDream V1, or another HiDream model for your goal. Image models create images, while video models support text-to-video and image-to-video.
Enter a text description or upload image and video references. Describe the subject, scene, action, shot, style, sound, and narrative.
Adjust aspect ratio, video duration, resolution, visual style, and other parameters. HiDream V1 supports 5–20 second 1080p videos.
HiDream understands intent, plans scenes, shots, actions, and sound, then generates the corresponding image or video.
If the result needs adjustment, edit the prompt or references and generate again, then export the final work.
Each model direction serves a different creative task. Available models and parameters are shown in the creation workspace.
| Model | Focus | Best for |
|---|---|---|
| HiDream-O1-Image | Image generation and editing | AI images, text rendering, subject consistency |
| HiDream-I1 | Open-source image generative foundation model | Text-to-image generation |
| HiDream-E1 | Instruction-based image editing | Input-image editing with instructions |
| HiDream V1 | Native omnimodal video | Text-to-video, image-to-video, audio-visual sync |
| HiDream-O1-World | Interactive world model | Digital scene building and interactive worlds |
| HiDream-O1-Embodied | Embodied intelligence | Visual understanding and physical action |
Find answers to common questions about AI image editing with HiDream and its advanced intelligent transformation features.
HiDream is an advanced image generation and editing model from Google, designed to deliver clearer visual quality, stronger reference-image fidelity and more precise fine-tuning. It produces more natural lighting and textures, maintains subject consistency more reliably and follows complex visual instructions more accurately across multiple edits.
Unlike traditional image editors, HiDream combines intelligent prompt understanding with consistent character editing and superior scene preservation. Our AI maintains character identity throughout the editing process while ensuring realistic style transformations and reliable multi-character adjustments.
HiDream apps handle focused production tasks such as removing backgrounds and objects and resizing creative for different formats. Explore all available apps
Yes. Upload your picture in Edit Image, highlight the area you want changed, and type the change. Look over the whole picture before you post it, because any AI edit can change something you did not ask for.
Yes! The intuitive natural-language interface of HiDream makes professional image editing accessible to everyone. Beginners can achieve stunning results simply by describing their vision, while professionals can use advanced features for complex editing workflows.
Thanks to our optimized AI architecture, HiDream delivers fast image transformations. Most edits are completed in seconds, enabling rapid iteration and efficient creative workflows without compromising quality.
Yes, HiDream focuses on high-resolution image rendering and superior scene preservation. Our AI ensures edited images maintain their original quality while seamlessly integrating new elements and transformations.
Absolutely. HiDream excels at reliable multi-character adjustments, maintaining character consistency in complex scenes. Our AI can edit multiple people or objects while preserving their individual features and relationships.
We are here to help you make the most of the AI image editing capabilities of HiDream. For questions, feedback or technical support, contact us at [email protected]. Our professional team will respond within 1-2 business days to ensure you get the best editing experience.