Head to head: Krea 3 Edit vs Wan
Krea 3 Edit vs Wan
By Ryan Merket · Published
Krea 3 Edit and Wan trade strengths across image-editing briefs that test exact counts, typography, spatial relationships, perspective, and material detail. The results hinge less on broad image quality than on which model follows the prompt’s most specific constraints.
Krea 3 Edit takes the matchup, 62.8 to 60.0, with an limited confidence verdict and a 5–2 edge in task wins, plus one tie. That’s a consistent advantage, not a blowout: the aggregate gap is modest, but Krea won more of the varied briefs. Wan’s wins were concrete. It delivered the rainy claymation harbor scene with its requested props and atmosphere, and it got the exact-count task right with seven cups where Krea produced eight. Those are meaningful successes on prompts where a single missed constraint can sink the result. Krea was stronger across the design and composition tests. It made the plum-soda poster’s wording more prominent, put the desk under the window, and handled the bakery alley’s centered vanishing point and hand-painted look more convincingly. Its library aisle also had more natural perspective and scale; on the reflections brief, it better showed the requested objects in the chrome, though neither image fully nailed the water-surface reflection requirement. The quince-tart comparison was a tie: Wan made the fruit, sugar, and pastry layers clearer, while Krea delivered the more convincing macro photograph. Across the full set, Wan’s standout wins don’t outweigh Krea’s broader reliability on layout, perspective, and design-specific instructions. **Verdict: Krea 3 Edit wins—clearly on the task record, if narrowly on aggregate score.**
Harbor tea stall without cups
At blue hour on a misty wharf, depict a cozy claymation-style tea stall called Brine & Thyme serving steaming saffron kettles and hanging paper lanterns, with a sleepy tabby cat on a stool, coils of rope, gull feathers, and wet wooden planks reflecting amber light — but the entire scene must contain NO cups, mugs, glasses, or teacups anywhere; cinematic 16:9 composition, gentle drizzle, tactile stop-motion textures, richly detailed but uncluttered.
OpenAI: GPT-6 Luna: Model B makes the claymation styling, gentle drizzle, steaming kettles, lanterns, cat, ropes, feathers, and wet reflections especially clear, while keeping the scene free of cups or drinking glasses. Model A has a stronger misty, cinematic feel and an excellent sign, but its drizzle and stop-motion texture are less evident. (Second judge pass, order swapped — scores are the average of both: Model B captures the rainy, lantern-lit wharf with steaming kettles, a sleepy tabby, rope, feathers, and w | Anthropic: Claude Fable 5.1: Model B delivers the requested claymation/stop-…
Exact counting
A clean overhead flat lay on white marble of EXACTLY seven identical espresso cups arranged in a single straight row, each clearly separate and individually countable, even soft lighting, 16:9.
OpenAI: GPT-6 Luna: Model B shows exactly seven clearly separated cups in a straight row on white marble, while Model A shows eight cups. Model A has a clean, balanced layout, but Model B adheres more closely to the exact-count requirement and also includes espresso in the cups. (Second judge pass, order swapped — scores are the average of both: Model B shows exactly seven espresso cups in a clear straight row on white marble, while Model A shows eight cups and they appear empty. Model A is cleanly arranged, but Mo | Anthropic: Claude Fable 5.1: Model A shows eight cups, failing the exact-cou…
Macro quince tart slice
An extreme macro close-up of a single slice of quince tart on a silver dessert fork, photographed with a 100mm macro lens aesthetic: razor-sharp detail on the glossy quince glaze, sugar crystals, and flaky pastry layers, while the background falls into creamy shallow bokeh showing only soft hints of a mint-green tablecloth and blurred candlelight; elegant editorial food photography, side-lit, intimate composition, 16:9.
OpenAI: GPT-6 Luna: Model B more clearly shows the glossy quince, distinct sugar crystals, flaky pastry layers, and silver fork, with the mint-green setting and candlelight softly blurred behind it. Model A has appealing macro detail and bokeh, but its pastry and fork feel less natural and the slice is less clearly presented. (Second judge pass, order swapped — scores are the average of both: Model B more clearly delivers the intimate macro food-photo brief, with glossy quince, visible sugar crystals, flaky pastry, | Anthropic: Claude Fable 5.1: Model A delivers a more convincing macro photog…
Readable plum soda poster
Design a crisp Art Nouveau poster pinned to a café wall advertising a sparkling drink, featuring an illustrated bottle, plum blossoms, curling bubbles, and a cream background with deep indigo and copper accents; the text must be fully legible and easy to read, exactly as follows: "FIZZLING PLUM" on the main line and "Batch 07" beneath it, with clean typography, balanced spacing, and warm afternoon window light in a 16:9 frame.
OpenAI: GPT-6 Luna: Model A presents the exact text prominently on one main line and combines the bottle, blossoms, curling bubbles, and warm café light in a richly detailed Art Nouveau design. Model B is clean and fully framed, but splits the main text across two lines and feels less like a sparkling-drink poster. (Second judge pass, order swapped — scores are the average of both: Model A more fully captures the Art Nouveau brief, with the exact wording prominently arranged on one main line and a vivid bottle, blo | Anthropic: Claude Fable 5.1: Model A delivers authentic Art Nouveau letterin…
Spatial layout
A clean isometric illustration of a bedroom: a bed against the LEFT wall, a round rug centered on the floor, a desk under the WINDOW on the back wall, and a floor lamp in the FRONT-RIGHT corner. Flat-vector style, consistent perspective.
OpenAI: GPT-6 Luna: Model A places the desk directly beneath the back-wall window and follows the requested layout with a centered round rug and a front-right floor lamp. Model B is a crisp flat-vector illustration, but its desk sits beside rather than under the window, and its patterned background is more distracting. (Second judge pass, order swapped — scores are the average of both: Model A places the desk directly beneath the window and more clearly satisfies the requested layout, while Model B places the desk | Anthropic: Claude Fable 5.1: Model A places every element correctly—bed flush…
Bakery alley vanishing point
A hand-painted gouache illustration of a narrow old-town bakery alley viewed from eye level, with a strong single vanishing point centered at the far end: in the foreground a giant wicker basket of oversized sesame loaves, midground stacked crates of pears and bottled lemonade outside a tiny shop named Crumb Relay, and in the distance small pedestrians under striped awnings; all objects must obey believable relative scale and perspective, with straight architectural lines, soft morning sun, and a balanced 16:9 composition.
OpenAI: GPT-6 Luna: Model A better captures the hand-painted gouache style and a centered, eye-level alley vanishing point while including the oversized bread basket, pears, lemonade, pedestrians, awnings, and correctly rendered shop name. Model B includes most requested details, but its more graphic style and less centered perspective make the scene feel less aligned with the prompt. (Second judge pass, order swapped — scores are the average of both: Model A better captures the hand-painted gouache style and cente | Anthropic: Claude Fable 5.1: Model A delivers a true centered vanishing poin…
Perspective & scale
A photorealistic one-point-perspective shot down a long empty library aisle, shelves receding to a single vanishing point, books getting consistently smaller with distance and nothing warped or wrongly sized, warm overhead light, 16:9.
OpenAI: GPT-6 Luna: Model A delivers a convincing, symmetrical one-point-perspective aisle with consistent recession, warm overhead lighting, and a photorealistic finish. Model B is bright and orderly, but its shelf geometry and book scale feel less natural and the image reads more like a render. (Second judge pass, order swapped — scores are the average of both: Model B delivers the centered, warm-lit library aisle, while Model A has a stronger, more convincing one-point perspective with clearer depth and consiste | Anthropic: Claude Fable 5.1: Model A delivers a convincing photorealistic on…
Reflections & glass
A photorealistic close-up of a chrome teapot on a polished dark marble counter next to a glass of water; the window and a red apple on the counter must be correctly reflected in BOTH the chrome and the water's surface, dramatic side light, 16:9.
OpenAI: GPT-6 Luna: Model A better matches the close-up framing and dramatic side light, with the window and red apple more evident in the chrome reflections. Model B is cleaner and balanced, but neither image clearly shows both the window and apple reflected in the water’s surface as requested. (Second judge pass, order swapped — scores are the average of both: Model A more clearly shows the window and a red apple reflected in the chrome, with dramatic side lighting; its glass also has more legible reflections at | Anthropic: Claude Fable 5.1: Model A delivers the harder, more dramatic side…
Matchup powered by OpenRouter.