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 12 shared challenges, with an AI judge weighing in on each.
HiDream I1 Full
#60 of 62 in Text-to-Image
Stable Diffusion 3.5 Large Turbo
#61 of 62 in Text-to-Image
Where the votes landed
HiDream I1 Full
0.0%
win rate
Ties
0.0%
Stable Diffusion 3.5 Large Turbo
100.0%
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 photographic realism and natural lighting.
- + Perfectly followed the spatial instruction for the red book sitting on top of the cube.
- + High-quality reflections and refraction within the glass cube.
- − The blue sphere is levitating rather than resting on the bottom of the cube.
Stable Diffusion 3.5 Large Turbo
- + Clean, sharp rendering with vibrant colors.
- + Accurately placed the plant behind the glass cube.
- − Failed the core spatial instruction: the red book is inside the cube instead of on top.
- − The composition feels more like a 3D product render than a natural photograph.
Verdict: HiDream I1 Full followed all spatial instructions accurately, correctly placing the red book on top of the glass cube and providing a very realistic photographic aesthetic. Stable Diffusion 3.5 Large Turbo failed the prompt's logic by placing the book inside the cube, resulting in a less realistic and less accurate image.
Candid Street Photography
Text-to-Image“A candid street photo of an elderly Japanese man repairing a red bicycle in light rain, reflections on wet pavement, shallow depth of field, 50mm lens, natural skin texture, imperfect framing, motion blur from passing cars, cinematic but realistic, no stylization.”
AI Judge Analysis
HiDream I1 Full
- + Excellent handling of wet pavement reflections and light rain atmosphere
- + Convinced realistic skin textures and believable lighting
- + Captures the cinematic, shallow depth of field look requested
- − The structural integrity of the bicycle is warped (missing seat post connection)
- − Lacks the requested 'motion blur' for the passing cars in the background
Stable Diffusion 3.5 Large Turbo
- + Successfully applied the requested 'imperfect framing' with a tighter, lower crop
- + Clean colors and sharp subject
- − Very poor anatomical rendering of the hands and face
- − The rain effect and wet pavement reflections are weak and unrealistic
- − The bicycle geometry is nonsense with multiple overlapping frames
Verdict: HiDream I1 Full produces a much more visually pleasing and realistic image that captures the 'cinematic but realistic' atmosphere of a rainy street. While the bicycle in HiDream I1 Full has structural issues, Stable Diffusion 3.5 Large Turbo fails significantly on character anatomy and misses the 'natural skin texture' and 'shallow depth of field' requirements entirely.
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 'beads in braids' and 'battle-worn' descriptors.
- + Stronger atmospheric lighting with realistic reflections on the metal plate armor.
- + Highly detailed facial texture including realistic skin pores and varied scarring.
- − The composition is very centered and traditional.
- − Slightly over-sharpened digital look in the hair texture.
Stable Diffusion 3.5 Large Turbo
- + Natural, dynamic composition and head tilt.
- + Good engraving details on the armor plates.
- − Missed the 'small beads' in the hair prompt entirely.
- − The 'scars' look more like fresh red paint or blood smudges rather than battle-worn scars.
- − Skin texture is overly smooth and lacks the requested lifelike detail and dirt.
Verdict: HiDream I1 Full followed the detailed prompts much more effectively, specifically including the beads in the braids and creating a convincing battle-worn appearance. Stable Diffusion 3.5 Large Turbo produced a cleaner, more stylized image but failed to capture the 'worn' aesthetic and the specific hair details requested.
Modern Clean Menu
Text-to-Image“Modern minimalist restaurant menu design, white background with colorful food photos in grid, sections for appetizers/pizza/mains, bold sans-serif fonts, vibrant accents, clean professional layout for casual dining.”
AI Judge Analysis
HiDream I1 Full
- + Successfully integrated text sections directly with corresponding food imagery.
- + Maintains a clean, professional white background that looks like a real menu.
- + Food photography is realistic and high quality.
- − Text is mostly gibberish despite having a legible structure.
- − Limited variety in food, showing only pizza for different sections.
Stable Diffusion 3.5 Large Turbo
- + Features a vibrant grid of diverse food items.
- + Better variety of food colors and shapes.
- + Includes distinct sections for 'Pizza' and 'Mians'.
- − The layout is cluttered and looks more like a collage than a menu design.
- − Food items look CGI or plastic-like rather than professional photography.
- − Includes repetitive, nonsensical text strings in the lists.
Verdict: HiDream I1 Full is the superior choice for this task as it actually produces a coherent 'menu design' that aligns with modern minimalist standards. While Stable Diffusion 3.5 Large Turbo provides more colorful food, its layout is chaotic and the food renders look artificial and unappetizing for a restaurant setting.
Chalkboard Menu
Text-to-Image“Handwritten-style chalkboard menu in a cozy café, all text rendered in the exact same realistic chalk handwriting style with natural variations in letter size, slight slant, and chalk texture — no printed or digital fonts anywhere on the board. Title at the top in elegant cursive chalk handwriting: ‘TODAY’S SPECIALS – APRIL 30, 2026’. Below it, three menu items also in the same handwritten chalk style: ‘Truffle Mushroom Risotto – $24’, ‘Grilled Octopus with Lemon & Herbs – $28’, ‘Brown Butter Chocolate Chip Cookies – $9’. At the very bottom, smaller text in the identical handwritten chalk style (slightly smaller but still clearly legible with the same handwriting characteristics): ‘All items made fresh daily • Ask about our gluten-free options’. Warm ambient café lighting, visible chalk dust and smudges, realistic handwriting imperfections, no clean printed text anywhere.”
AI Judge Analysis
HiDream I1 Full
- + The chalk texture looks more authentic to a real blackboard.
- + Correctly rendered the specific date requested: 'April 30, 2026'.
- + The title cursive is elegant and matches the requested style.
- − Completely failed to render the specific menu items requested, using nonsensical filler text instead.
- − The text style looks more like a digital font than natural handwriting in several places.
- − Includes a strange 'Fart Sant' menu item and nonsensical pricing.
Stable Diffusion 3.5 Large Turbo
- + Successfully attempted the specific menu items like 'mushroom risotto' and 'grilled octopus'.
- + The environment and composition are more visually appealing as a 'cozy café'.
- + The handwriting has more natural variations and thickness characteristic of real chalk.
- − Failed to render the date correctly, showing 'apri 3L 20-'.
- − The spelling of menu items is inconsistent and contains errors like 'ocotpgs' and 'risctto'.
- − The overall text rendering is messy and difficult to read in some sections.
Verdict: Both models struggled with the complex text requirements. HiDream I1 Full succeeded in rendering the correct date and title, but completely ignored the specific menu items in favor of gibberish. Stable Diffusion 3.5 Large Turbo attempted the full prompt including specific food items, but failed on the date and suffered from significant spelling errors throughout.
The Reversed Rodeo
Text-to-Image“Horse riding astronaut in space — horse on top, not vice versa. Surreal, highly detailed, cinematic.”
AI Judge Analysis
HiDream I1 Full
- + Excellent cinematic lighting and atmosphere
- + High level of texture detail on the astronaut suit and horse's coat
- + Dynamic and natural pose for both horse and rider
- − The horse legs appear to be emerging from or merging into smoke/clouds at the bottom
- − Traditional composition which, while beautiful, feels less surreal than requested
Stable Diffusion 3.5 Large Turbo
- + Strong adherence to the surreal requirement with a floating, dream-like quality
- + Clean, high-contrast visual style
- + Interesting perspective and layout
- − The horse features anatomical glitches, particularly the oddly shaped and positioned rear legs
- − The astronaut is facing the opposite direction of the horse's travel, creating a disjointed feel
- − Lower detail level compared to Image A
Verdict: While both models failed the specific trick prompt instruction to place the 'horse on top' of the astronaut, HiDream I1 Full is the superior image due to its exceptional cinematic quality and realistic textures. Stable Diffusion 3.5 Large Turbo provides a more surreal atmosphere, but it is marred by significant anatomical distortions in the horse's legs.
The Capybara Taxi Driver
Text-to-Image“Photorealistic scene inside a yellow New York taxi at night. A capybara is driving, wearing a yellow taxi driver cap and a dark jacket. It has a calm, professional expression and both front paws on the steering wheel. In the back seat sits a human businesswoman in a coat, looking at her phone with a completely normal, bored expression (as if this is just another normal ride). Through the windows you can see the streets of Manhattan at night with blurred lights. Realistic taxi interior, photorealistic, detailed fur and fabric, 35mm lens, night lighting with reflections, shallow depth of field.”
AI Judge Analysis
HiDream I1 Full
- + Excellent adherence to the 'human looking at phone' prompt requirement
- + High level of detail in the capybara's fur and expression
- + Clear, legible text on the taxi cap
- − The capybara's paws look more like human hands or paws with too many joints
- − The passenger's hand/vortex near the phone is slightly messy
Stable Diffusion 3.5 Large Turbo
- + Strong cinematic lighting reflecting a night scene
- + Aesthetically pleasing bokeh in the background
- + Good capybara anatomy for the paws
- − Completely missed the passenger's action; she is looking forward instead of at a phone
- − The passenger's face is very blurry and lacks the requested bored expression
- − The text on the hat is distorted
Verdict: HiDream I1 Full is the clear winner as it followed every detail of the prompt, including the specific behavior of the passenger looking at her phone. While Stable Diffusion 3.5 Large Turbo has more atmospheric lighting, it failed to render the passenger correctly and missed the core narrative element of the scene.
The Halloween Invitation
Text-to-Image“Vintage gothic Halloween party invitation. Dark parchment poster, spooky border with webs and thorns, central glowing jack-o-lantern, bats, twisted trees, moody night sky. Add elegant gothic title text saying "Halloween Party Invitation", a small scroll banner saying "You are invited to a night of frights", and event details at the bottom: Date: 30.10.2026 Time: 7pm Location: The Arches, NYC Spooky but polished, cinematic lighting, square format.”
AI Judge Analysis
HiDream I1 Full
- + Strong composition with a clear central focus and decorative scroll banner.
- + Captured the gothic aesthetic well with the twisted trees and cemetery background.
- + Successful reproduction of the 'You are invited...' phrase on the banner.
- − Failed significantly on the specific event details, inventing 'Yootmber' and incorrect times.
- − Text at the bottom becomes illegible gibberish.
Stable Diffusion 3.5 Large Turbo
- + Excellent thorny border and web details as requested in the prompt.
- + Clean, modern-gothic font for the main title.
- − Completely missed the secondary banner text and all specific event details (Date, Time, Location).
- − The composition feels empty in the middle and bottom sections compared to the prompt requirements.
Verdict: HiDream I1 Full followed the layout and complexity of the prompt much better, including the requested banner and background elements, though it hallucinated the specific date and time text. Stable Diffusion 3.5 Large Turbo failed to include more than half of the requested text elements, leaving the invitation incomplete. HiDream I1 Full is the winner for capturing the structure and most of the text content, despite some spelling errors.
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 'JAPAN' and 'SUSHI' correctly spelled and positioned.
- + Accurate adherence to the isometric perspective and solid light blue background.
- + High-quality 3D render style with clean, refined textures for the sushi and base.
- − Missing the requested flag icon.
- − The wasabi/ginger garnish is very minimal and slightly lacks detail.
Stable Diffusion 3.5 Large Turbo
- + Creative use of a physical sign for the 'JAPAN' text.
- + Includes a flag icon as requested in the prompt.
- + Good 3D miniature diorama feel with realistic wood and ceramic textures.
- − Spelling error in 'SUSHI' which is rendered as 'SIIHI'.
- − Text is not at the top-center as requested.
- − The flag icon does not correctly represent the Japanese flag.
Verdict: HiDream I1 Full is the clear winner for its superior text rendering and adherence to the layout instructions, despite missing the small flag icon. Stable Diffusion 3.5 Large Turbo features a spelling error ('SIIHI') and fails to place the text in the requested 'top-center' position, instead integrating it into the scene as props.
Adorable Baby Animals in Sunny Meadow
Text-to-Image“Hyper-photorealistic scene of fluffy baby animals—a golden retriever puppy, tabby kitten, baby bunny, and red fox kit—with big expressive eyes and ultra-detailed soft fur, playfully chasing butterflies and tumbling together in a lush wildflower meadow, warm golden sunrise light with god rays and dew sparkles, joyful wholesome vibe, 8K masterpiece.”
AI Judge Analysis
HiDream I1 Full
- + Excellent variety of animals that mostly align with the prompt requirements.
- + Beautiful warm lighting and convincing 'god rays' effects.
- + Sharp details on the fur and expressive, clear eyes.
- − Missing the baby bunny mentioned in the prompt, rendering two kittens instead.
- − The composition is a bit static for a 'tumbling' and 'chasing' prompt.
Stable Diffusion 3.5 Large Turbo
- + Dynamic 'tumbling' composition with a playful sense of motion.
- + Vibrant backlighting with nice high-contrast highlights on the fur.
- − Fails significant parts of the prompt, missing the bunny and fox entirely.
- − Anatomical issues where the cat's limbs appear to merge or sprout awkwardly from the center.
- − The dog's breed looks more like an Australian Shepherd mix than a Golden Retriever.
Verdict: HiDream I1 Full is the clear winner despite missing the bunny, as it successfully renders a variety of recognizable baby animals with high-quality fur and lighting. Stable Diffusion 3.5 Large Turbo fails to include the fox or bunny and suffers from anatomical incoherence where the animals overlap.
Vintage Cafe Logo
Text-to-Image“Vintage minimalist restaurant logo for "Caffè Florian", retro cloche dome with steam and "Est. 1720" banner, classic typography, warm brown and cream tones, subtle texture on light background, vector emblem style.”
AI Judge Analysis
HiDream I1 Full
- + Clean vector aesthetic suitable for a modern logo
- + Accurate text rendering for the 'Est. 1720' segment
- + Follows the brown and cream color scheme well
- − Completely missed the main brand name 'Caffè Florian'
- − The composition puts a coffee cup inside the cloche which looks slightly cluttered
- − Lacks the requested 'subtle texture' on the background
Stable Diffusion 3.5 Large Turbo
- + Successfully includes both the brand name and the establishment date
- + Strong interpretation of vintage style with professional shading and texture
- + Clever integration of the cloche dome onto a coffee mug shape
- − Minor spelling error in the brand name ('Caffeé' instead of 'Caffè')
- − Slightly less 'minimalist' than requested due to high-contrast shading
Verdict: Stable Diffusion 3.5 Large Turbo is the clear winner as it actually followed the primary instruction to include the restaurant name 'Caffè Florian', whereas HiDream I1 Full omitted it entirely. Although Stable Diffusion had a minor character error in the name, its overall vintage aesthetic, texture, and creative composition far outshine the overly simplified and incomplete output from HiDream.
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 with readable labels
- + Adheres well to the requested color palette
- + Clear vector art style with minimal gradients
- − Confused layout logic that does not follow the 6-step sequence
- − Included literal prompt text like '+ ICON' in the labels
- − Contains icons that look more like space shuttles than the Saturn V or Lunar Module
Stable Diffusion 3.5 Large Turbo
- + Sophisticated infographic layout with a clear hierarchy
- + Excellent stylistic consistency for a poster design
- + Captures the 'NASA-inspired' aesthetic very effectively
- − Text is mostly gibberish/lorem ipsum
- − Fails to follow the specific 6-step sequence requested in the prompt
- − Does not include the specific icons requested for each stage
Verdict: HiDream I1 Full followed the text requirements more literally, but suffered from logical layout issues and the inclusion of prompt keywords in the final design. Stable Diffusion 3.5 Large Turbo produced a much more professional-looking infographic that felt like a real poster, though it failed to correctly list the six specific mission steps requested. HiDream I1 Full is the likely winner for better adherence to the specific content requested, despite the text errors.
Explore each model
Distilled version of SD 3.5 Large that generates high-quality images in just 4 steps, offering faster inference and reduced costs