HiDream AI's image-to-image editing model for instruction-based image modifications and transformations
Settled by community votes across 4 shared challenges, with an AI judge weighing in on each.
HiDream E1
0.7
arena score
#33 of 32 in Image Editing
Skill signature
· Image Editing
HiDream I1 Full
13.7
arena score
#60 of 62 in Text-to-Image
Vote tally
Where the votes landed
HiDream E1
0%
win rate
Ties
0%
HiDream I1 Full
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 E1
HiDream I1 Full
AI Judge Analysis
HiDream E1
- + Successfully added a full head of hair as requested.
- + Maintains the facial identity and expression of the subject reasonably well.
- + Preserves the lighting and much of the original jacket's structure.
- − The hair has a synthetic, overly wavy texture that looks like a wig.
- − The background has been completely altered from a rocky desert to a sandy, blurry landscape.
- − The skin texture on the forehead and face has become blotchy and unrealistic.
HiDream I1 Full
- + The added hair texture in the back looks realistic and matches the color.
- − Completely failed the main prompt by leaving the top of the head bald.
- − Added bizarre hallucinated objects like a black strap around the beard and a large backpack.
- − Completely replaced the background and significantly altered the person's facial features and eyes.
Verdict: HiDream E1 followed the instructions for a full head of hair, though it failed to preserve the original background and the hair texture looks somewhat artificial. HiDream I1 Full failed the core task, keeping the subject bald while hallucinating strange accessories and completely changing the background and facial details. HiDream E1 is the clear winner for actually performing the requested edit.
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 E1
HiDream I1 Full
AI Judge Analysis
HiDream E1
- + Successfully incorporates all elements: tv anchor blazer, dog, and a hockey rink setting with players.
- + Distinctive caricature style with exaggerated facial features.
- + Good storytelling by placing the character in a broadcast-style environment.
- − The facial likeness to the source image is weak.
- − Significant anatomy issues with the hands, including an extra finger on the right hand.
HiDream I1 Full
- + Maintains a much stronger facial likeness to the woman in the source image.
- + Preserves the denim jacket style from the original photo.
- + Clean illustration style with fewer grotesque anatomical artifacts.
- − Fails to include any clear hockey references as requested.
- − The 'tv show anchor' profession is only vaguely implied by the background TV, lacking the specific professional attire or setting seen in HiDream E1.
Verdict: HiDream E1 followed the complex prompt more thoroughly by including the hockey and tv anchor elements, though it lost the likeness of the subject and suffered from poor hand anatomy. HiDream I1 Full created a much better caricature of the specific person, but completely ignored the hockey requirement and provided a weaker interpretation of the professional setting. HiDream E1 is the likely winner for better adherence to the multifaceted editing instructions despite the technical flaws.
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 E1
HiDream I1 Full
AI Judge Analysis
HiDream E1
- + Excellent preservation of the original pose and character layout
- + Maintains the distinct facial expressions of the 'distracted boyfriend' meme
- + Adheres to the pastel color palette requested
- − Art style is more generic 'webcomic' than the specific hand-painted Ghibli aesthetic
- − Loss of clothing details like the plaid pattern on the shirt
HiDream I1 Full
- + Stronger adherence to the Studio Ghibli artistic style and textures
- + Background feels more 'lived-in' and detailed in line with Ghibli's hand-painted look
- + Adds charming character details like the hat and floral patterns
- − Completely changes the identity and specific expressions of the people
- − The 'annoyed' expression of the girlfriend is lost in favor of a neutral look
- − The man is smiling instead of making the iconic 'ooh' face
Verdict: HiDream E1 prioritizes the core structure and narrative of the source image, ensuring the 'Distracted Boyfriend' meme is still recognizable despite the stylistic shift. HiDream I1 Full creates a much more convincing Ghibli-style illustration with beautiful textures, but it loses the specific character expressions that make the original image iconic.
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 E1
HiDream I1 Full
AI Judge Analysis
HiDream E1
- + Excellent preservation of the original subjects (woman and dog).
- + Successfully added wind effect to the hair while keeping the face intact.
- + Followed all instructions including the addition of flying leaves.
- − The added leaves look like flat, artificial stickers.
- − The lighting on the leaves does not match the scene.
HiDream I1 Full
- + Successfully captured a high-energy, lively atmosphere.
- + The motion of the hair and leaves feels more integrated into the scene's lighting.
- − Completely failed to preserve the source image, removing the dog entirely.
- − Drastically changed the identity and pose of the woman.
- − Anatomical issues with the hands appearing distorted.
Verdict: HiDream E1 is the clear winner as it successfully performed the edit while maintaining the integrity of the original image, keeping the woman and her dog consistent. HiDream I1 Full generated a brand new image that ignored the source content, most notably removing the dog which was a central part of the request's context.
Next steps
Explore each model
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