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 8 shared challenges, with an AI judge weighing in on each.
HiDream I1 Full
#60 of 62 in Text-to-Image
Reve Image 1.0
#42 of 62 in Text-to-Image
Where the votes landed
HiDream I1 Full
0%
win rate
Ties
0%
Reve Image 1.0
0%
win rate
Challenge by challenge
The strongest take from each model on every shared challenge, with the AI judge's read.
Geometric Composition
Text-to-Image“A glass cube on a wooden table. Inside the cube is a small blue sphere. On top of the cube sits a red book. A green plant is behind the cube, partially visible through the glass. Soft window light from the left.”
AI Judge Analysis
HiDream I1 Full
- + Excellent handling of glass transparency and reflections
- + High textural detail on the book and table
- + Realistic soft lighting from the window side
- − The plant is more alongside the cube than directly behind it
- − The sphere appears to be levitating without a clear physical resting point
Reve Image 1.0
- + Natural wood grain on the table
- + Good spatial arrangement of requested objects
- + Clean, modern aesthetic
- − The sphere is floating unnaturally in the center of the glass
- − Reflections in the glass are slightly simplified
- − The red book looks somewhat like a box or a very thick notebook due to simplified edges
Verdict: Both models followed the prompt instructions accurately, including all four requested items and the specific lighting direction. HiDream I1 Full is the winner due to its superior rendering of glass materials and caustic light patterns on the table, which feel more grounded and realistic compared to the slightly more digital look of Reve Image 1.0.
Fantasy Warrior
Text-to-Image“Close portrait of a battle-worn paladin in ornate engraved plate armor, hair braided with small beads, faint scars and dirt on the skin, warm torchlight reflecting off metal, shallow depth of field, bokeh sparks, lifelike eyes, highly detailed texture on leather straps and cloth underlayer.”
AI Judge Analysis
HiDream I1 Full
- + Excellent adherence to the 'battle-worn' description with prominent scars and dirt.
- + Intricate details on the engraved plate armor and leather strap buckles.
- + Strong symmetrical composition with well-balanced warm lighting.
- − The scars appear somewhat stylized and less than 'faint'.
- − The eyes have a slightly over-sharpened, artificial quality.
Reve Image 1.0
- + Beautifully soft, realistic lighting with natural-looking bokeh sparks.
- + Highly realistic skin texture and very lifelike, convincing eyes.
- + High-quality engraving detail on the armor surfaces.
- − The subject appears quite young, which may slightly clash with the 'battle-worn' descriptor.
- − The composition is off-center with less visual impact than Model A.
Verdict: HiDream I1 Full delivers a more dramatic and iconic 'paladin' look with better framing and stronger adherence to the battle-worn aesthetic. Reve Image 1.0 has superior realistic textures and lighting, but the choice of a younger subject and less engaging composition makes it less effective as a character portrait for this specific prompt.
Bald man challenge
Image Editing“Give the person a full, thick head of natural hair with realistic texture, density, and a natural hairline. Preserve facial features and lighting.”
AI Judge Analysis
HiDream I1 Full
- + Attempts to add hair behind the head.
- − Completely fails to preserve the source image, changing the person's identity and the entire background.
- − Adds nonsensical artifacts like the strange black object across the face.
- − Top of the head remains bald, directly contradicting the prompt.
Reve Image 1.0
- + Excellent source preservation, keeping the person's face, glasses, clothing, and background perfectly intact.
- + Successfully adds a full, thick head of hair as requested.
- + Maintains original lighting and texture consistency.
- − The hairline on the forehead is slightly soft and could be more defined.
Verdict: Reve Image 1.0 successfully performed the requested edit with high fidelity, adding thick hair while perfectly preserving the original person and environment. HiDream I1 Full failed on almost every metric, generating a new image with a different face, background, and bizarre artifacts while leaving the character mostly bald.
Isometric Miniature Diorama Scenes
Text-to-Image“Create a clear, 45° top-down isometric miniature 3D cartoon scene of Japan's signature dish: sushi, with soft refined textures, realistic PBR materials, gentle lighting, on a small raised diorama base with minimal garnish and plate. Solid light blue background. At top-center: 'JAPAN' in large bold text, 'SUSHI' below it, small flag icon. Perfectly centered, ultra-clean, high-clarity, square format.”
AI Judge Analysis
HiDream I1 Full
- + Excellent text rendering with clean, professional typography.
- + Higher level of detail in the 3D sushi models including rice textures and roe.
- + Very high clarity and adherence to the 'refined' texture prompt.
- − Failed to include the requested small flag icon.
- − The top sushi roll is slightly merging with the back edge of the display.
Reve Image 1.0
- + Successfully included the requested flag icon.
- + Strong adherence to the isometric perspective and diorama base requirement.
- + Clean, minimalist cartoon style that matches the prompt.
- − Text is rendered in a casual handwritten style rather than 'large bold text'.
- − The 3D models are very simplistic, lacking the 'refined textures' and 'PBR material' quality of the competitor.
- − The green sushi piece looks unnatural and low-detail compared to the others.
Verdict: HiDream I1 Full produces a much higher quality 3D render with impressive textures and professional-grade typography, though it missed the flag icon. Reve Image 1.0 followed more of the prompt instructions (including the flag) but the visual quality is significantly lower, appearing more like a simple mobile game asset than a refined 3D diorama. HiDream I1 Full is the likely winner due to its superior visual fidelity and clean execution.
Over-the-top cartoon caricature
Editing“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.”
AI Judge Analysis
HiDream I1 Full
- + Successfully captures a caricature art style while keeping the person recognizable.
- + Includes the dog prominent in the foreground.
- + Background details (TV and figures) subtly hint at the hockey and TV job.
- − Fails to clearly integrate hockey elements into the main character's outfit or props.
- − The caricature is less 'humorous' and more of a standard portrait illustration.
- − Anatomy of the hand/arm on the shoulder is confusing.
Reve Image 1.0
- + Excellent adherence to all prompt elements: TV anchor desk, microphone, hockey stick, hockey jersey, and dog.
- + The dog wearing a tie and headset adds a great touch of humor.
- + Strong, vibrant composition that clearly communicates the requested persona.
- − The facial likeness is slightly softened compared to the source image.
- − The hockey stick is floating/attached oddly to the desk.
Verdict: Reve Image 1.0 is the clear winner as it fully incorporated every aspect of the prompt (hockey, TV anchor, and dogs) into a cohesive and humorous scene. HiDream I1 Full created a high-quality caricature, but it missed the overt 'professional' cues requested, leaving the hockey element as a small, blurry detail on a background screen.
Studio Ghibli Anime Style
Editing“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”
AI Judge Analysis
HiDream I1 Full
- + Excellent character design in a high-quality anime style
- + Vibrant colors and cheerful mood
- + High resolution and clean linework
- − Fails to preserve the core narrative of the source image (the 'distracted boyfriend' meme structure)
- − Changes the character's expressions to be happy rather than stunned/shocked
Reve Image 1.0
- + Perfectly preserves the composition and emotional narrative of the source image
- + Captures the hand-painted, soft texture characteristic of Studio Ghibli backgrounds
- + Maintains the specific clothing patterns (the plaid shirt) while transforming the style
- − The foreground character's face is quite blurry
- − Anatomy on the rightmost character's arm is slightly awkward
Verdict: While HiDream I1 Full produces a high-quality illustration, it fails the basic task of preserving the source image's context, turning the famous 'distracted boyfriend' meme into a generic happy group. Reve Image 1.0 successfully transforms the entire scene into the requested Studio Ghibli aesthetic while perfectly maintaining the specific poses, expressions, and identifiable characteristics of the original photo.
Golden Hour Stroll
Image Editing“Add dynamic motion to this photo: make hair blow in the wind, add leaves flying, energetic and lively feel.”
AI Judge Analysis
HiDream I1 Full
- + Excellent sense of motion and energy
- + Successfully adds wind-blown hair and flying leaves
- − Fails at source preservation by completely changing the person's face and identity
- − Removes the dog from the image
- − Changes the background and overall composition significantly
Reve Image 1.0
- + Perfectly preserves the source image subject, dog, and background
- + Accurately applies all requested edits including motion hair and flying leaves
- + High-quality integration of new elements into the existing scene
- − Some leaves in the foreground have slightly unnatural motion blur compared to others
Verdict: Reve Image 1.0 is the clear winner as it successfully followed the editing instructions while perfectly preserving the original subject, dog, and background. HiDream I1 Full failed the core task of 'editing' by generating a completely new image that changed the person's identity and removed the dog entirely.
Apollo 11: Journey to Tranquility
Text-to-Image“Create a clean, modern vector infographic poster about the Apollo 11 mission. NASA-inspired palette (navy, white, muted red, light gray). Flat-vector style, crisp lines, consistent iconography, subtle gradients only. Steps (stop at landing): 1. Launch (Saturn Vicon) 2. Earth Orbit (Earth + orbit ring icon) 3. Translunar (trajectory arc icon) 4. Lunar Orbit (Moon + orbit ring icon) 5. Descent (lunar module descending icon) 6. Landing (lunar module on the surface icon) Small supporting elements (minimal text): • Crew strip: three silhouette icons with only last names: Armstrong, Aldrin, Collins. • Landing site marker: Moon pin labeled "Tranquility" only. Layout constraints: generous margins, large readable labels, clean background with subtle stars. Vector-only, print-poster look, high resolution.”
AI Judge Analysis
HiDream I1 Full
- + Strong typography and visual impact with a bold title.
- + Excellent adherence to the color palette and flat-vector style.
- + High-quality, clean vector illustrations with crisp lines.
- − Failed to provide all 6 specific steps, combining icons haphazardly.
- − Text includes nonsensical additions like '+ UNG' and '+ ICON'.
- − Includes icons of a space shuttle and a ringed planet (Saturn) which are historically inaccurate for Apollo 11.
Reve Image 1.0
- + Followed the multi-step structural instruction much more accurately.
- + Included specific crew names (Armstrong, Aldrin, Collins).
- + Correctly depicted the lunar module specifically for 'Descent' and 'Landing' phases.
- − Multiple spelling errors in the main headers and subtitles ('LUNNR', 'DESSENT', 'APQUO').
- − Visual density is a bit low with a lot of empty dark blue space.
- − Icons for the lunar module are somewhat cluttered and less 'clean' than Model A's vectors.
Verdict: Reve Image 1.0 is the winner despite several typos because it followed the complex structural requirements of the prompt, including the specific 6-step progression and the lunar module details. HiDream I1 Full produced a much cleaner and more aesthetically pleasing poster, but it failed the core instruction by providing generic space icons (like a space shuttle) and only showing four mismatched phases.
Explore each model
Reve AI's text-to-image generation model with strong aesthetic quality, accurate text rendering, and detailed instruction following capabilities