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
#11 of 62 in Text-to-Image
Where the votes landed
GPT Image 2
100.0%
win rate
Ties
0.0%
Recraft V4
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
GPT Image 2
- + Excellent adherence to lighting instructions with clear soft window light from the left.
- + Highly realistic textures on the wooden table and red book cover.
- + Clean, logical geometry for the glass cube and the blue sphere's placement.
- − The sphere appears more like a matte ball than a glass or polished marble, though this is a minor stylistic detail.
Recraft V4
- + Successfully includes all requested elements including the green plant visible through the glass.
- + The glass cube has realistic thickness and internal reflections.
- − Physical logic is flawed as the blue sphere appears to be levitating inside the cube.
- − The perspective of the red book is slightly awkward, appearing very flat against the top of the cube.
- − Visible digital artifacts or 'noise' on the wooden table surface.
Verdict: GPT Image 2 is the superior image due to its superior lighting, realistic textures, and grounded physics. While Recraft V4 included all prompt elements, it failed to place the sphere logicially on the bottom of the cube, resulting in an unintentional levitation effect.
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 skin texture and hyper-realistic facial details
- + Clean and authentic mechanical details on the bicycle and tools
- + Strong adherence to the imperfect framing prompt with the foreground post and skewed depth
- − The sense of rain is very subtle and barely visible on the ground
- − The background motion blur feels more like static bokeh than moving cars
Recraft V4
- + Excellent atmospheric rendering of rain and wet pavement reflections
- + Great interpretation of motion blur from passing cars
- + Strong cinematic composition and lighting
- − The man's anatomical proportions and posture are slightly awkward
- − The bicycle has structural issues including a disconnected front fork/axle area
Verdict: GPT Image 2 (Model A) wins on realism and technical detail, particularly in the rendering of the man and the bicycle's hardware. While Recraft V4 (Model B) nails the atmospheric effects like the rain and motion blur better, it suffers from structural glitches in the bicycle and a slightly less realistic human figure.
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 photographic realism with natural skin texture and lifelike eyes.
- + Very detailed engraving and weathering on the plate armor.
- + Beautifully soft boken and lighting that matches the requested warm torchlight atmosphere.
- − The 'small beads' in the hair are present but less distinct than in the other model.
- − The battle-worn aspect is subtle, looking more like freckles than significant dirt/scars.
Recraft V4
- + Perfect adherence to specific details like the hair beads and leather strap buckles.
- + The expression and skin texture convey a much stronger 'battle-worn' and scarred feel.
- + Excellent composition with clear bokeh sparks and visible torchlight sources.
- − The lighting on the face is slightly flat compared to the dramatic lighting on the armor.
- − The armor engraving, while good, looks a bit more generic/digital than Image A.
Verdict: Both models followed the prompt exceptionally well, but Recraft V4 captured more of the specific 'battle-worn' grit and detailed accessories like the hair beads and leather straps requested. While GPT Image 2 has a superior high-end photographic aesthetic, Recraft V4 feels more like the specific character described in the prompt.
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
- + Exceptional text rendering with perfect legibility and branding.
- + Highly professional layout with sophisticated graphic design elements.
- + Complete adherence to all prompt sections including pricing and descriptions.
- − The 'Sirloin Steak' image shows some slight texture blurriness compared to the other photos.
Recraft V4
- + Clean minimalist aesthetic with consistent isolated food photography.
- + Excellent spatial balance in the 3x3 grid layout.
- + Correct spelling and prices rendered in a clear sans-serif font.
- − Lacks the 'vibrant accents' and branding elements requested.
- − A bit too sterile, missing the 'bold' professional flourishes seen in model A.
Verdict: GPT Image 2 (Model A) is the clear winner as it produces a fully realized, professional-grade restaurant menu with branding, icons, and detailed descriptions that look ready for print. While Recraft V4 (Model B) followed the grid and section instructions well, its output feels like a basic template compared to the rich, commercial-quality design of GPT Image 2.
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 text rendering with impressive fiery glow effects.
- + Highly detailed and appetizing 'exploded' burger composition.
- + Perfect adherence to all requested elements including the starburst pricing.
- − The composition is a bit crowded with the text and burger overlapping.
Recraft V4
- + Clean, more traditional advertisement layout.
- + Realistic lighting on the food items.
- − Failed to apply the fiery glowing effect to the secondary text and price.
- − The starburst element is very basic and lacks the requested glowing effect.
- − The lettuce and patty textures appear slightly less crisp compared to Image A.
Verdict: GPT Image 2 followed the prompt's stylistic requirements much more closely, specifically regarding the fiery, glowing treatment of all text elements and the starburst. Recraft V4 produced a clean image but ignored the glow effect for the majority of the text and price, making it look less cohesive as a 'Magic Burger' ad.
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 with realistic grit and smudge effects.
- + Perfect completion of the truncated prompt text with 'Brown Butter Chocolate Chip Cookies'.
- + Highly realistic cursive title and natural variations in handwriting.
- − The slant of the text lines is slightly inconsistent.
Recraft V4
- + Perfect text legibility and alignment.
- + Atmospheric café background provides good context.
- + Accurately completed the menu items and pricing.
- − The text looks a bit too much like a digital chalk-style font rather than organic handwriting.
- − The 'cursive' title requirement was not fully met, as it uses a semi-print style.
Verdict: GPT Image 2 (Model A) is the clear winner because it achieved a significantly more authentic chalk-on-blackboard texture, complete with realistic variations in pressure and slanting. While Recraft V4 produced a very clean and readable image, the lettering feels more like a digital font overlay rather than the handwritten style requested in the prompt.
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
- + Perfectly adheres to the specific inversion instruction of the horse being on top.
- + Excellent textures on the spacesuit and realistic lighting reflections on the visor.
- + Highly creative and humorous interpretation of a difficult prompt.
- − The horse's front legs and hoof shape are anatomically awkward.
- − The rein/harness placement doesn't logically connect to the horse's mouth.
Recraft V4
- + Stunning cinematic composition with great use of scale and atmospheric debris.
- + High level of detail in the horse's muscular structure and the astronaut's suit.
- + Beautiful lighting and color palette.
- − Completely failed the negative constraint/inversion: the astronaut is on top, not the horse.
- − The horse appears to have five legs visible in the dust cloud.
Verdict: GPT Image 2 is the clear winner because it successfully followed the complex instruction to have the horse riding the astronaut, whereas Recraft V4 ignored the specific 'horse on top' requirement. While Recraft V4 achieved a more aesthetically pleasing cinematic style, its failure to adhere to the core prompt logic makes it an incorrect output for this specific challenge.
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 photorealistic texture on the capybara's fur
- + Accurate depiction of a New York taxi partition and interior details
- + Passenger has a perfect bored expression as requested
- − The capybara only has one visible paw on the wheel instead of two
Recraft V4
- + Creative wide-angle composition showing more of the taxi environment
- + Accurately places both paws on the steering wheel
- + Realistic raining window effect adds to the atmosphere
- − The scale of the capybara is slightly too small for a driver's seat
- − The passenger's shoes are disproportionately large and chunky
- − The 'TAXI' sign is positioned strangely on the interior ceiling line
Verdict: GPT Image 2 (Model A) is the winner because it achieves a much higher level of photorealism and better anatomical consistency for the human passenger. While Recraft V4 (Model B) followed the 'two paws' instruction more literally, its overall image quality is marred by warped proportions and a less convincing taxi interior.
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
- + Perfect execution of the vintage parchment aesthetic with weathered textures.
- + Exquisite typography that integrates seamlessly into the overall gothic graphic design.
- + Excellent adherence to all layout elements including the scroll banner and detailed border.
- − The jack-o-lantern face is slightly generic, though it fits the theme well.
Recraft V4
- + Good use of cobweb borders as requested in the prompt.
- + Correct text rendering for all the specific event details.
- + Dynamic sense of movement with the bats and rain-like atmosphere.
- − The 'scroll banner' is just a modern ribbon shape that lacks a vintage aesthetic.
- − The composition feels much less like a 'dark parchment poster' and more like a standard digital illustration.
- − The transition between the image and the black border is somewhat harsh.
Verdict: GPT Image 2 is the clear winner as it masterfully captures the requested 'vintage gothic' aesthetic with sophisticated textures and professionally integrated typography. While Recraft V4 includes all the correct text, its final output lacks the cinematic lighting and cohesive parchment feel provided by GPT Image 2.
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 3D miniature diorama execution with high-quality PBR textures on the fish and wood.
- + Perfect text rendering and alignment according to the prompt.
- + Rich, appealing composition that fits the 'cartoon scene' descriptor perfectly.
- − Includes a bit more garnish than requested by the 'minimal garnish' instruction.
Recraft V4
- + Follows the 'minimal garnish' instruction well with a simple plate setup.
- + Good lighting and shadows that give a realistic sense of depth.
- − The diorama base appears to be made of crushed ice/salt which is less 'clean' than the prompt implied.
- − The sushi modeling is slightly less refined with some odd proportions on the maki roll.
- − Typography is thin and loses the 'bold' requirement from the prompt.
Verdict: GPT Image 2 perfectly captures the requested 3D cartoon aesthetic with bold, clear typography and a professionally finished diorama look. While Recraft V4 produces a more realistic texture for the food, it fails to meet the 'bold text' requirement and its diorama base is more cluttered and less stylistically cohesive than the clean presentation in GPT Image 2.
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 character interaction and dynamic posing expressing a 'tumbling' motion.
- + Warm, vibrant color palette that perfectly captures the golden sunrise lighting.
- + High clarity and realistic textures on the fur and butterflies.
- − The fox kit has slightly unusual paw anatomy on its leading leg.
- − The rabbit is relegated to the background, appearing less active than the others.
Recraft V4
- + Successfully includes all requested elements with a sense of forward momentum.
- + The butterflies are integrated well into both the foreground and background for depth.
- + Atmospheric lighting with clear sunbeam effects through the trees.
- − The kitten has slightly distorted, crossed eyes that look less natural.
- − The rabbit is positioned awkwardly in the center as if floating or standing on its hind legs.
- − General fur texture is slightly softer and less defined than in model_a.
Verdict: Both models captured the heartwarming prompt well, but GPT Image 2 stands out for its superior character posing and richer color saturation. Recraft V4 accurately depicts the scene and god rays but suffers from minor anatomical issues in the animals' faces and a less cohesive composition.
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
- + Perfectly followed the request for a banner element for the established date.
- + Exhibits a beautiful woodcut/engraving texture that fits the vintage aesthetic.
- + Superior typography with appropriate spacing and professional flourishes.
- − The composition is quite complex for a 'minimalist' logo request.
Recraft V4
- + Follows the minimalist part of the prompt more closely with a cleaner design.
- + Good use of negative space in the cloche and steam illustration.
- + Correct spelling and accent placement on the text.
- − Failed to include the requested banner for the 'Est. 1720' text.
- − Total absence of the 'subtle texture' requested in the prompt, appearing very flat.
- − The font choice for 'Florian' feels more like 1970s retro than classic 1700s vintage.
Verdict: GPT Image 2 followed the prompt much more accurately by including the specific banner element and the requested texture. While Recraft V4 captured the 'minimalist' aspect well, it missed several key descriptive requirements and used a typography style that felt modern-retro rather than classical. GPT Image 2's woodcut style and balanced composition make it the clear winner for a vintage heritage brand.
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 text rendering throughout.
- + Highly cohesive professional layout with clear sequential numbering.
- + Rich detail in the icons while maintaining a clear infographic style.
- − Includes more gradients and 3D lighting than requested in the 'flat-vector' style prompt.
- − Some minor perspective issues on the lunar module's legs in the final landing step.
Recraft V4
- + Adheres much more strictly to the 'flat-vector' and 'crisp lines' request.
- + Accurately uses the requested NASA-inspired color palette.
- + Logical layout of the requested six steps and crew details.
- − Composition feels sparse and lacks the 'poster' feel requested.
- − The 'Saturn V' icon is a generic rocket silhouette rather than a specific Saturn V representation.
- − Text alignment and hierarchy are somewhat basic.
Verdict: GPT Image 2 (Model A) produces a much more finished and professional poster with impressive text rendering and a comprehensive layout, though it leans toward a semi-realistic illustrative style rather than pure flat vector. Recraft V4 (Model B) follows the stylistic constraints of 'flat vector' more literally but fails to create a compelling or balanced composition. Overall, GPT Image 2 is the preferred output due to its superior visual impact and completeness as an information graphic.
Explore each model
Recraft's latest text-to-image generation model with high-quality output, supporting various aspect ratios and custom color palettes