Black Forest Labs' 12-billion parameter multimodal flow transformer for in-context image generation and editing with character consistency, typography handling, and commercial-ready quality
Settled by community votes across 4 shared challenges, with an AI judge weighing in on each.
FLUX.1 Kontext [pro]
20.3
arena score
#41 of 62 in Text-to-Image
Skill signature
· Text-to-Image
HiDream E1
0.7
arena score
#33 of 32 in Image Editing
Vote tally
Where the votes landed
FLUX.1 Kontext [pro]
0%
win rate
Ties
0%
HiDream E1
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.”
FLUX.1 Kontext [pro]
HiDream E1
AI Judge Analysis
FLUX.1 Kontext [pro]
- + Excellent preservation of the source image's background, lighting, and textures.
- + Highly realistic hair texture and a believable natural hairline.
- + Maintains the subject's identity and facial features with high fidelity.
- − The hair styling is slightly modern compared to the rugged aesthetic of the source.
HiDream E1
- + Successfully added a large volume of hair as requested.
- − Completely altered the background to a generic desert landscape.
- − The hair texture appears synthetic and plastic-like rather than natural.
- − Significantly changed the lighting and skin texture of the subject's face.
Verdict: FLUX.1 Kontext [pro] performed an excellent image edit, seamlessly integrating a realistic head of hair while preserving every other detail of the source image. In contrast, HiDream E1 failed the preservation criteria by generating a new background and applying an artificial, doll-like hair texture that does not match the original scene's lighting or quality.
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
FLUX.1 Kontext [pro]
HiDream E1
AI Judge Analysis
FLUX.1 Kontext [pro]
- + Excellent preservation of the subject's clothing and basic facial features from the source image.
- + Captures the caricature style with expressive features and a vibrant comic-book aesthetic.
- − Completely failed to incorporate the specific profession (TV anchor), dogs, or hockey themes from the prompt.
HiDream E1
- + Successfully incorporates all prompt elements, including the TV anchor setting, a dog, and hockey players.
- + High-quality caricature style with exaggerated features that still retain a resemblance to the original subject.
- − Loss of source image garment consistency (swapped denim for a red blazer).
- − Anatomical issues in the hands and the dog's legs.
Verdict: FLUX.1 Kontext [pro] creates a great stylistic caricature of the person in the photo but ignores more than half of the prompt's instructions regarding the setting and hobbies. HiDream E1, while changing the subject's clothes, successfully builds a scene that includes the TV anchor role, dog, and hockey players as requested, making it a much better execution of the specific editorial request.
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
FLUX.1 Kontext [pro]
HiDream E1
AI Judge Analysis
FLUX.1 Kontext [pro]
- + Excellent adherence to the Studio Ghibli art style with watercolor-like textures.
- + Preserved the specific blue plaid pattern of the man's shirt from the source image.
- + Captures the facial expressions and character positions with high accuracy to the original meme.
- − The woman in the foreground has a somewhat generic anime face compared to the more distinct Ghibli character designs.
HiDream E1
- + Successfully applied a warm, nostalgic color palette.
- + Maintains the composition and character placement of the source image.
- − Failed to preserve the blue plaid shirt, changing it to a solid pink shirt.
- − The art style is more akin to generic vector web-art than the requested hand-painted Ghibli style.
- − Loss of detail in facial expressions, especially the jilted girlfriend's look of indignation.
Verdict: FLUX.1 Kontext [pro] is the clear winner as it successfully interpreted the 'Studio Ghibli' request with beautiful hand-painted textures and soft line work while maintaining key details like the man's plaid shirt. HiDream E1 produced a much flatter, generic illustration style and failed to preserve the clothing details from the source image.
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.”
FLUX.1 Kontext [pro]
HiDream E1
AI Judge Analysis
FLUX.1 Kontext [pro]
- + Excellent source preservation, maintaining the woman's face and dog's appearance perfectly.
- + Subtle and realistic blowing hair effect.
- + Natural integration of flying leaves that match the environment's color palette.
- − The 'energy' is relatively low as the pose remains identical to the static source image.
HiDream E1
- + Provides a very dynamic hair-blowing effect.
- + Strong adherence to the request for 'energetic' feel by adding significant motion.
- − Poor source preservation; the woman's face is noticeably changed and looks more digital/artificial.
- − The added leaves are an unrealistic red color and have a 'stamped-on' appearance.
- − The dog's face and fur details are degraded compared to the original.
Verdict: FLUX.1 Kontext [pro] is the superior model because it successfully added wind and leaves while perfectly preserving the identity of the woman and the dog from the source image. In contrast, HiDream E1 significantly altered the woman's facial features and added unrealistic, high-contrast leaves that clash with the original lighting and color scheme.
Next steps
Explore each model
HiDream AI's image-to-image editing model for instruction-based image modifications and transformations