HiDream AI's 17B parameter text-to-image model using sparse diffusion transformer with mixture of experts, achieving state-of-the-art image generation quality with strong prompt following
Settled by community votes across 4 shared challenges, with an AI judge weighing in on each.
HiDream I1 Full
15.2
arena score
#60 of 62 in Text-to-Image
Skill signature
· Text-to-Image
Qwen Image Edit 2511
20.2
arena score
#16 of 32 in Image Editing
Vote tally
Where the votes landed
HiDream I1 Full
0%
win rate
Ties
0%
Qwen Image Edit 2511
0%
win rate
Shared challenges
4
Challenge by challenge
The strongest take from each model on every shared challenge, with the AI judge's read.
Bald man challenge
Image EditingEdit instruction
“Give the person a full, thick head of natural hair with realistic texture, density, and a natural hairline. Preserve facial features and lighting.”
HiDream I1 Full
Qwen Image Edit 2511
AI Judge Analysis
HiDream I1 Full
- + Attempts to add hair length at the back.
- − Completely changes the subject's face and identity.
- − Fails to add hair to the top of the head as requested.
- − Introduces a bizarre black artifact/mask over the beard.
- − Changes the background and lighting entirely.
Qwen Image Edit 2511
- + Excellent adherence to the request for a full head of hair with realistic texture.
- + Perfectly preserves the subject's original facial features, glasses, and expression.
- + Seamlessly maintains the original background, lighting, and clothing.
- − The hairline on the forehead is slightly high, though still natural-looking.
Verdict: Qwen Image Edit 2511 followed the instructions perfectly, adding a realistic and well-integrated head of hair while keeping the person's identity and the background intact. HiDream I1 Full failed significantly, generating a different person entirely, changing the setting, and adding strange visual artifacts.
Over-the-top cartoon caricature
EditingEdit instruction
“Create a caricature of me and my job. Make it exaggerated and humorous, incorporating my profession as a tv show anchor and my love for dogs and hockey.”
Source
HiDream I1 Full
Qwen Image Edit 2511
AI Judge Analysis
HiDream I1 Full
- + Successfully captures the subject's likeness in a caricature style.
- + Includes a subtle hockey-related image on the TV in the background.
- + Incorporates a dog as requested.
- − The 'TV anchor' profession is poorly represented, as the outfit and living room setting feel casual.
- − The hockey reference is too small and backgrounded to be a significant part of the theme.
- − The dog has a strange black appendage coming out of its head near the woman's mouth.
Qwen Image Edit 2511
- + Clearly represents the 'TV anchor' profession with a studio desk, suit, and microphone.
- + Includes a prominent dog character that matches the caricature style.
- + High energy and humorous exaggeration suitable for the prompt.
- − Completely misses the 'hockey' requirement.
- − Likeness is less accurate to the source image compared to the other model.
- − The hand holding the microphone and the hand with the pen have anatomical errors.
Verdict: Both models failed to fully incorporate all elements of the prompt, particularly the hockey requirement. HiDream I1 Full maintained a much better likeness to the original woman and included a small hockey reference, but failed to make her look like a TV anchor. Qwen Image Edit 2511 captured the anchor profession perfectly but ignored the hockey aspect and struggled with hand anatomy.
Studio Ghibli Anime Style
EditingEdit instruction
“Transform this photo into a Studio Ghibli–inspired illustration. Use soft pastel colors, hand-painted textures, gentle lighting, dreamy backgrounds, and a warm, nostalgic mood”
Source
HiDream I1 Full
Qwen Image Edit 2511
AI Judge Analysis
HiDream I1 Full
- + Captures the Ghibli art style effectively through character designs and line work.
- + Creates a vibrant, themed environment with nostalgic elements.
- + Maintains the core narrative pose of the original meme.
- − Low source preservation, completely redrawing characters and changing their clothing.
- − Loses the specific facial expressions that define the 'Distracted Boyfriend' meme.
- − Introduces extra background characters that clutter the composition.
Qwen Image Edit 2511
- + Excellent source preservation, keeping the original faces, poses, and clothing patterns recognizable.
- + Applies a soft, painterly texture that aligns well with the request for hand-painted feel.
- + Maintains the crucial emotional storytelling of the source image while stylizing it.
- − The art style is more generic 'digital art' than specific Ghibli-esque character design.
- − Slightly less 'dreamy' background compared to model A.
Verdict: HiDream I1 Full produces a more authentic Ghibli animation aesthetic but fails as an editor by completely replacing the original people and clothes. Qwen Image Edit 2511 succeeds as an image editing tool by beautifully stylizing the source photo with the requested textures and lighting while leaving the identity and expressions of the subjects perfectly intact.
Golden Hour Stroll
Image EditingEdit instruction
“Add dynamic motion to this photo: make hair blow in the wind, add leaves flying, energetic and lively feel.”
HiDream I1 Full
Qwen Image Edit 2511
AI Judge Analysis
HiDream I1 Full
- + Strong sense of energy and motion through active posing
- + Successfully adds a large number of falling leaves
- − Fails to preserve the source image subject, changing her face, clothing, and removal of the dog
- − Poor source preservation in terms of background details and composition
Qwen Image Edit 2511
- + Excellent source preservation, keeping the original person, dog, and background intact
- + Effectively applies the requested edits to the hair and environment while maintaining the original's integrity
- − The wind effect on the hair is slightly stylized and looks a bit less natural than the rest of the scene
Verdict: Qwen Image Edit 2511 is the clear winner as it successfully follows the editing instructions while maintaining the identity of the person, the presence of the dog, and the overall composition of the source image. HiDream I1 Full fails as an image editor, essentially generating a new image that resembles the prompt but ignores the visual data of the source photo, such as the specific face and the dog.
Next steps
Explore each model
Alibaba's Qwen image editing model for instruction-based image modifications and transformations