Black Forest Labs' state-of-the-art image generation model with maximum quality and speed, supporting text-to-image and multi-reference image editing with up to 4MP output
Settled by community votes across 4 shared challenges, with an AI judge weighing in on each.
FLUX.2 [pro]
26.3
arena score
#8 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.2 [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.2 [pro]
HiDream E1
AI Judge Analysis
FLUX.2 [pro]
- + Excellent source preservation, keeping the background and clothing nearly identical.
- + Hair matches the beard texture and the original lighting perfectly.
- + Maintains the original facial features and skin texture with high fidelity.
- − The hairline is slightly high, giving a slightly receding look rather than a 'full' head of hair.
HiDream E1
- + Successfully adds a very thick and dense head of hair.
- + Matches the hair color to the existing beard.
- − Fails to preserve the original background, changing the desert landscape significantly.
- − Alters the person's face significantly, adding extra wrinkles and changing the nose/eye shape.
- − The hair texture looks artificial and overly smooth compared to the rugged beard.
Verdict: FLUX.2 [pro] followed the instructions perfectly, providing a realistic head of hair while preserving every detail of the person, clothing, and background. HiDream E1 failed as an image editor, completely regenerating the background and altering the subject's face so much that it no longer resembles the source image.
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.2 [pro]
HiDream E1
AI Judge Analysis
FLUX.2 [pro]
- + Successfully incorporates all elements: tv anchor desk, hockey sticks/pucks, and multiple dogs.
- + Maintains character likeness while translating into a clean vector-style caricature.
- + Excellent composition with the hockey rink integrated into the anchor desk.
- − The style is more of a digital cartoon than a classic hand-drawn caricature.
HiDream E1
- + Applies a 3D Pixar-style caricature look with exaggerated facial features.
- + Includes the hockey rink background and a dog as requested.
- − Lost the subject's likeness, changing hair color and facial structure significantly.
- − The 'tv anchor' profession is less clear compared to the literal desk in the other model.
- − Anatomical issues with the hands holding the dog.
Verdict: FLUX.2 [pro] is the clear winner as it successfully synthesized all parts of the prompt—TV anchoring, hockey, and dogs—into a cohesive and well-composed scene while maintaining the subject's identity. HiDream E1 struggled with the subject's likeness and had significant artifacts in the hands and character details.
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.2 [pro]
HiDream E1
AI Judge Analysis
FLUX.2 [pro]
- + Excellent adherence to the 'hand-painted textures' and Ghibli style requests.
- + Preserves original character poses and details like the plaid shirt pattern perfectly.
- + Maintains the distinct facial expressions that are key to the original meme.
- − The watercolor border effect might be slightly distracting for some users.
HiDream E1
- + Successfully adopts a general anime aesthetic.
- + Features soft pastel colors as requested.
- − Loses significant character details such as the plaid pattern on the shirt.
- − The expressions and poses are weakened, losing the specific 'distracted boyfriend' meme energy.
- − The art style lacks the specific textural depth of a Ghibli illustration, appearing more like modern vector art.
Verdict: FLUX.2 [pro] is the clear winner as it perfectly captures the Studio Ghibli aesthetic with hand-painted textures and soft lighting while maintaining the integrity of the original meme's poses and expressions. In contrast, HiDream E1 produces a generic anime style that loses important details like the man's plaid shirt and the girlfriend's sharp look of betrayal.
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.2 [pro]
HiDream E1
AI Judge Analysis
FLUX.2 [pro]
- + Excellent source preservation, maintaining identical facial features and background details from the original.
- + Realistic wind effect on hair and clothing that feels integrated with the scene.
- + The flying leaves appear naturally blurred and textured, consistent with the environment's lighting.
- − The sheer number of leaves in the foreground can feel slightly overwhelming or distracting from the subject.
HiDream E1
- + Successfully applied the requested wind effect to the hair.
- + Good implementation of motion-blurred leaves in the background.
- − Failed to preserve the source image identity, noticeably altering the subject's face.
- − The colors are overly saturated compared to the original, losing the natural summer park feel.
- − The foreground leaves appear like flat, red stickers rather than integrated elements.
Verdict: FLUX.2 [pro] is the clear winner as it perfectly preserves the identity and background of the source image while adding high-quality, realistic motion effects. HiDream E1 failed the primary objective of image editing by significantly changing the subject's face and using stylized, unrealistic leaf graphics that clash with the original photo's aesthetic.
Next steps
Explore each model
HiDream AI's image-to-image editing model for instruction-based image modifications and transformations