OpenAI's state-of-the-art image generation model with arbitrary resolution up to 4K and strong instruction following
Settled by community votes across 13 shared challenges, with an AI judge weighing in on each.
GPT Image 2
#4 of 62 in Text-to-Image
Recraft V4.1
#53 of 62 in Text-to-Image
Where the votes landed
GPT Image 2
0.0%
win rate
Ties
0.0%
Recraft V4.1
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
GPT Image 2
- + Excellent adherence to the 'behind' placement of the plant.
- + Highly realistic glass reflections and refraction.
- + Perfect composition and lighting from the left as requested.
- − The glass cube looks more like a frame with thick glass edges rather than a solid or simple hollow cube.
Recraft V4.1
- + Beautiful material rendering on the blue sphere.
- + Sharp focus and high-quality textures on the wooden table.
- + Clean, solid appearance of the glass cube.
- − The blue sphere appears to be floating mid-air, which is physically inconsistent.
- − The plant is positioned to the side rather than 'behind' the cube as requested.
Verdict: GPT Image 2 followed the spatial instructions more accurately, placing the plant behind the cube so it is visible through the glass. While Recraft V4.1 has beautiful aesthetics and textures, the sphere is floating unnaturally and the plant placement is less precise compared to the prompt.
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
GPT Image 2
- + Excellent realism with authentic skin textures and clothing wrinkles
- + Accurately represents a Japanese street environment with legible signage and localized car styles
- + High level of detail in the bicycle and tool kit components
- − The 'light rain' is very subtle, with little visible falling water despite the wet ground
- − The 'motion blur' on the background car is slightly stiff looking
Recraft V4.1
- + Strong atmospheric mood with highly visible rain and beautiful reflections
- + Captures the 'candid' and 'imperfect framing' request more effectively than Model A
- + Excellent use of shallow depth of field and bokeh
- − Anatomical issues with the man's hands appearing mangled and fused
- − The bicycle structure is physically impossible with no chain or hub connection visible in the workspace
Verdict: GPT Image 2 is significantly more coherent and realistic, successfully depicting an authentic Japanese scene with perfect skin and object details, though it lacks a strong 'rain' atmosphere. Recraft V4.1 captures the requested mood and camera effects much better, but fails on a technical level with major anatomical distortions in the hands and broken geometry in the bicycle.
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
GPT Image 2
- + Excellent intricate engraving on the armor which looks weathered and practical.
- + The hair braiding with beads is beautifully integrated and realistic.
- + Superior resolution and clarity in the facial skin textures and eyes.
- − The lighting is a bit generic and lacks the dramatic 'warm torchlight' orange tones requested.
Recraft V4.1
- + Stronger adherence to the lighting prompt with obvious warm torchlight influence.
- + Captures the 'faint scars' prompt more explicitly with a clear forehead scar.
- + Good use of bokeh sparks in the foreground adding to the atmosphere.
- − The engraving on the armor looks a bit repetitive and less detailed than Image A.
- − The hair beads are present but look slightly more muddy in texture.
Verdict: Both models followed the prompt well, but GPT Image 2 produced a superior technical image with incredible detail in the armor engravings and skin texture. While Recraft V4.1 captured the specific 'scar' and 'lighting' atmospheric elements more literally, GPT Image 2 is the preferred output due to its higher overall visual quality and artistic composition.
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
GPT Image 2
- + Excellent typography and layout with perfect text legibility.
- + High-quality, appetizing food photography that looks professional.
- + Strong adherence to all prompt elements including specific sections and brand identity.
- − Some icons in the footer are slightly generic.
- − Slight text repetition in descriptions if read very closely.
Recraft V4.1
- + Clean, minimalist aesthetic that is very functional.
- + Accurate text rendering for all menu items and prices.
- + Effective use of white space.
- − The food photos are not in a cohesive grid as requested, but randomly spaced icons.
- − Much more basic layout that lacks the professional 'vibrant accents' and branding of a real menu.
- − Large empty gaps in the center make the design feel unfinished.
Verdict: GPT Image 2 (Model A) is the clear winner as it produces a fully-realized, professional restaurant menu with beautiful food photography and a logical layout. Recraft V4.1 (Model B) is technically accurate with its text but fails to create a cohesive 'grid' of photos or a design that feels like a finished marketing asset.
Magic Burger Explosion: Fiery Photorealism Challenge
Text-to-Image“Ad for 'Magic Burger'. Dynamic, exploded burger with all components (bun, patty, cheese, lettuce, tomato, sauce) suspended in mid-air. Emphasize photorealistic detail and a sense of motion. Dark, fiery background with glowing embers. Integrate text: 'MAGIC BURGER' as a prominent title, 'LIMITED TIME ONLY' as a secondary message, and '€6.99' in a starburst, all rendered with a fiery, glowing effect.”
AI Judge Analysis
GPT Image 2
- + Excellent adherence to text styles, creating a cohesive fiery-glow effect across all three requested text elements.
- + Superior photorealistic textures on the burger components, particularly the sesame bun and juicy patty.
- + Highly dynamic composition with sauce droplets and embers filling the frame to create a sense of motion.
- − The image is quite crowded, with the large text and starburst competing heavily for the viewer's attention.
Recraft V4.1
- + Clean, professional-looking layout with more breathing room for the central burger subjects.
- + Clear, legible typography for the main title and subtext.
- − Failed to apply the 'fiery, glowing effect' to the starburst and subheader text.
- − The burger components look flatter and less appetizing compared to the realistic lighting in the other model.
- − The 'starburst' element looks like a basic clipart shape and lacks integration with the scene.
Verdict: GPT Image 2 is the clear winner for its superior rendering of photorealistic food textures and its much better execution of the fiery text effects requested in the prompt. While Recraft V4.1 offers a cleaner layout, it misses several stylistic cues for the text and the burger ingredients lack the dynamic, high-end commercial quality present in GPT Image 2.
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
GPT Image 2
- + Excellent chalk texture on the individual strokes.
- + Highly polished and realistic cafe environment and composition.
- + Perfect adherence to the complex text prompt with no spelling errors.
- − The cursive title is relatively simple rather than being 'elegant' ornamentation.
Recraft V4.1
- + Spells all the requested text correctly and concisely.
- + Includes organic chalk smudges on the board for extra realism.
- + Distinct variation in letter sizes and slants as requested.
- − The title is not in cursive as requested by the prompt.
- − The 'f' in Truffle has a strange overlapping artifact.
- − The composition is a tight crop rather than a cozy café scene.
Verdict: GPT Image 2 followed the prompt more closely, notably by providing the title in a cursive script and setting the board within a beautifully rendered, cozy café environment. While Recraft V4.1 excelled at the specific text rendering, it failed to provide the requested cursive style for the header and offered a less interesting composition.
The Reversed Rodeo
Text-to-Image“Horse riding astronaut in space — horse on top, not vice versa. Surreal, highly detailed, cinematic.”
AI Judge Analysis
GPT Image 2
- + Excellent adherence to the specific prompt instruction to have the horse on top of the astronaut.
- + High degree of realism in textures and surface details.
- + Strong conceptual humor and surreal execution.
- − Slightly creepy or unsettling character design.
- − Reins are physically nonsensical as they appear to be held by the horse's hooves.
Recraft V4.1
- + Beautiful cinematic lighting and environment effects.
- + Dynamic composition with a sense of motion.
- + Clean rendering of the astronaut and horse.
- − Total failure to follow the specific 'horse on top' instruction.
- − Generic interpretation of a common surrealist prompt.
Verdict: GPT Image 2 (Model A) followed the complex and unusual prompt request perfectly, showing a horse literally riding an astronaut. Recraft V4.1 (Model B) failed the negative constraint and provided a standard image of an astronaut riding a horse, making Model A the clear winner despite Model B's attractive visual style.
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
GPT Image 2
- + Excellent photorealism with cinematic lighting and realistic fur textures
- + Perfectly captures the 'bored' expression of the passenger in the background
- + High-quality rendering of the capybara's professional expression and clothing
- − The capybara's 'front paws' look more like human-like furry hands or gloves
- − Composition is a bit tight on the subject, making the taxi interior less visible
Recraft V4.1
- + Naturalistic side-profile composition showing the full interior and street context
- + Capybara's paws look more like actual animal paws on the steering wheel
- + Very clear depiction of the passenger using her phone and the taxi exterior
- − The passenger's hand holding the phone has anatomical issues (long, distorted fingers)
- − Capybara is not wearing the requested dark jacket
Verdict: GPT Image 2 (Model A) delivers a much more cinematic and polished image with superior lighting and texture, successfully capturing the specific 'bored' mood of the passenger. Recraft V4.1 (Model B) has a better composition for storytelling and handled the paws more realistically, but it failed to include the jacket on the capybara and has noticeable anatomical glitches on the human's hand.
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
GPT Image 2
- + Excellent typography with perfect spelling and professional gothic spacing.
- + Complex and balanced composition featuring atmospheric light and intricate border details.
- + Successfully incorporates secondary landscape elements like the castle and city skyline.
- − The parchment texture is slightly repetitive in the corners.
Recraft V4.1
- + Very vibrant and moody orange glow from the jack-o-lantern.
- + Creative twisted tree branches that frame the central subject.
- − Text rendering is poor, with the title feeling squashed and the banner text being very small/blurry.
- − Layout at the bottom is clumsy, with the event details floating in a single line without labels.
- − The border appears slightly less refined and more repetitive than Model A.
Verdict: GPT Image 2 is the clear winner as it produces a professional, ready-to-use invitation with perfect text rendering and a sophisticated layout. While Recraft V4.1 has nice lighting on the pumpkin, it fails the typography aspect of the prompt, resulting in a cluttered title and unreadable details.
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
GPT Image 2
- + Excellent text rendering with clear bold lettering and a clean flag icon.
- + Highly detailed and appealing miniature scene with varied sushi types and a stone lantern.
- + Great use of textures, particularly the wood grain and the glossy finish on the roe.
- − The garnish and scene elements are slightly more complex than the 'minimal' request.
- − The chopsticks are floating or resting awkwardly near the edge.
Recraft V4.1
- + Adheres strictly to the 'minimal garnish' and 45-degree top-down isometric instruction.
- + Achieves a very clean, soft 3D cartoon aesthetic with gentle lighting.
- + Accurate text spelling and centered composition.
- − The text is not 'large bold' as requested, appearing quite thin.
- − The sushi variety is quite low, featuring only three very similar pieces.
- − The flag icon is placed to the side rather than following the top-center hierarchy as well as Image A.
Verdict: GPT Image 2 captured the 'miniature 3D cartoon' and 'large bold text' instructions much better, producing a visually rich and high-quality diorama. While Recraft V4.1 followed the 'minimal' constraint more closely, it lacked the impact and texture quality found in the GPT Image 2 output.
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
GPT Image 2
- + Excellent fur texture across all animals
- + Clear and vibrant god rays with beautiful backlighting
- + Perfect subject framing and balance
- − The fox's front right paw looks slightly warped
- − Some butterflies are blurry or lacks structural detail
Recraft V4.1
- + High level of dynamic interaction between the animals
- + Visible dew sparkles on the grass as requested
- + Detailed sky with atmospheric volumetric lighting
- − The golden retriever's head anatomy and mouth shape look slightly unnatural
- − The kitten is hovering in a physically impossible way
- − Anatomy issues on the fox's legs
Verdict: Both models followed the prompt well, but GPT Image 2 provided a much more coherent and aesthetically pleasing composition with superior anatomical accuracy. While Recraft V4.1 captured the 'dew sparkles' better, its animals suffer from floating postures and distorted facial features in the puppy.
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
GPT Image 2
- + Excellent typography with correct accents and stylish serif fonts
- + Rich, detailed illustrative style with beautiful stippling and shading
- + Sophisticated composition featuring a professional border and banner
- − May be considered too ornate for a 'minimalist' request
Recraft V4.1
- + Adheres better to the 'minimalist' instruction with a clean, vector look
- + Accurate text and banner placement
- + Good use of negative space for a logo application
- − The accent mark on 'CAFFÈ' is detached and too large
- − Graphic elements like the steam and cloche are very basic compared to the other model
- − Lacks the 'subtle texture' requested in the prompt
Verdict: GPT Image 2 (Model A) provides a high-quality, professional-looking emblem that perfectly captures the vintage aesthetic and classic typography, though it leans more decorative than minimalist. Recraft V4.1 (Model B) is more minimalist but suffers from awkward spacing in the typography and a less refined illustration style. Model A is the clear winner for its superior visual quality and attention to detail in the 'Est. 1720' banner.
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
GPT Image 2
- + Excellent typography and nearly perfect spelling across a complex layout.
- + Sophisticated composition that uses all requested iconographic steps in a logical flow.
- + High-quality illustrative details that maintain a cohesive professional infographic aesthetic.
- − Slightly more illustrative than 'flat-vector', drifting into detailed 3D rendering for the lunar module.
Recraft V4.1
- + Strictly adheres to the 'flat-vector' style and minimalist NASA-inspired palette.
- + Clean, modern iconography that fits a traditional vector infographic look.
- + Accurately represents the crew and landing site in a simplified stylistic manner.
- − Contains numbering errors in the sequence labels (two '04's and no '05').
- − The Saturn V icon is very simplified and lacks the iconic proportions of the actual rocket.
- − The layout feels sparse with significant dead space in the bottom half.
Verdict: GPT Image 2 is the superior infographic, featuring professional-grade layout, excellent text rendering, and a comprehensive interpretation of all six mission steps. While Recraft V4.1 captures the 'flat-vector' aesthetic more accurately, it suffers from numbering errors and a less engaging composition compared to the densely detailed and informative NASA-themed poster produced by GPT Image 2.
Explore each model
Recraft's V4.1 standard tier text-to-image model — refines V4's photorealism with more natural lighting, softer gradients, and sharper illustration styles for everyday creative work