Head to head: Wan vs OpenAI: GPT-5.4 Image 2
Wan vs OpenAI: GPT-5.4 Image 2
By Ryan Merket · Published
A matchup of format discipline and exact scene control against stronger atmosphere and motion. Across eight varied image tasks, the models split their wins evenly, with distinct trade-offs in framing and execution.
Wan’s edge was consistency with the brief’s structure: it won perspective and scale, reflections and glass, bound colors and materials, and exact counting. Its images more often honored the requested 16:9 framing, and its seven-cup row was especially clear. OpenAI: GPT-5.4 Image 2 won legible text, Midnight Pantry Quartet, Tea Towel Whip, and Rainy Citrus Poster, with stronger atmosphere and more vivid action in several of those scenes. The aggregate scores were **58.5 for Wan** and **64.2 for OpenAI: GPT-5.4 Image 2**, but the saved result is a tie: four task wins apiece, with no task ties. The higher aggregate score doesn’t change that verdict. **Final call: Too close to call (not statistical).** Wan was the more dependable choice for requested framing and cleanly specified objects; OpenAI: GPT-5.4 Image 2 made several scenes feel more cinematic. This assessment applies to the saved task sample, not image generation in general.
Legible multi-line text
A minimalist event poster with three lines of crisp, correctly-spelled text stacked and centered: 'NIGHT MARKET' large on top, 'Fridays · 6–11pm' in the middle, 'Riverside Pier 4' at the bottom, on a deep navy background, clean sans-serif, subtle grain.
OpenAI: GPT-6 Luna: Model A renders all three lines crisply and correctly in a clean, centered hierarchy on a deep navy background with subtle grain. Model B also includes the correct text and strong centering, but its heavily textured, condensed headline is less crisp and less faithful to the requested clean sans-serif style. (Second judge pass, order swapped — scores are the average of both: Model B captures the poster feel especially well with a deep navy, subtly grained background and a balanced centered text h | Anthropic: Claude Fable 5.1: Both render all three lines correctly spelled a…
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 better matches the requested 16:9 format and delivers a warm, strongly centered aisle with shelves converging cleanly to one vanishing point. Model B is more conventionally photorealistic, but its square framing misses the requested aspect ratio and its lighting feels cooler and more institutional. (Second judge pass, order swapped — scores are the average of both: Model A better matches the requested 16:9 framing and presents a strong, balanced one-point perspective with consistently re | Anthropic: Claude Fable 5.1: Model A delivers the requested 16:9 frame, a cl…
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 requested 16:9 close-up and clearly shows the window and apple reflected in the chrome, while Model B has especially crisp reflections but is square rather than widescreen. Neither image clearly shows the apple reflected on the water’s surface, though both show window reflections there. (Second judge pass, order swapped — scores are the average of both: Model A better matches the requested 16:9 framing and shows the apple distinctly reflected in the chrome, while Model | Anthropic: Claude Fable 5.1: Model A delivers the requested 16:9 frame, dram…
Bound Colors Materials
A precise studio still life on a pale gray shelf featuring exactly five objects with strict attribute binding: a matte cobalt-blue ceramic teapot, a translucent amber glass vase, a brushed copper metal ladle, a soft sage-green linen napkin, and a glossy ivory plastic lemon squeezer; each object clearly separated, colors locked to the correct materials with no swapping or bleeding, neutral daylight, sharp focus, minimal shadows, front three-quarter composition, 16:9
OpenAI: GPT-6 Luna: Model A presents all five correctly bound objects in a clean, well-separated 16:9 studio composition. Model B also depicts the requested materials and colors convincingly, but its nearly square framing and more crowded object placement are less faithful to the prompt. (Second judge pass, order swapped — scores are the average of both: Model A better matches the requested 16:9 framing while clearly presenting all five objects with their colors and materials correctly bound. Model B is especially | Anthropic: Claude Fable 5.1: Model A delivers the requested 16:9 frame with a…
Midnight Pantry Quartet
A coherent multi-subject editorial scene of four household objects interacting like a tiny nocturnal gathering on a pantry shelf: a ribbed glass oil bottle leaning toward a striped ceramic salt cellar, a brass kitchen timer perched on a stack of recipe cards, and a pear balanced in a small saucer below them; all arranged in a balanced triangular composition with believable spatial overlap, soft moonlight from the left and a warm under-cabinet glow from the right, richly observed textures, 16:9
OpenAI: GPT-6 Luna: Model B creates a more convincing nocturnal gathering, with all the requested objects, rich textures, and clear moonlight-versus-warm-glow lighting; its square format misses the requested 16:9 framing. Model A matches the wide format and includes the objects, but its cards appear awkwardly supported by the pear and the scene feels less atmospheric. (Second judge pass, order swapped — scores are the average of both: Model B captures the nocturnal gathering with the requested moonlight, warm glow, | Anthropic: Claude Fable 5.1: Model A honors the 16:9 format and includes eve…
Tea Towel Whip
A sunlit kitchen action photograph of a mustard-yellow tea towel being snapped through the air by an unseen hand, caught at the exact instant it whips a spray of flour off a walnut countertop around a rolling pin, a blue enamel mug, and three tumbling apricots; dynamic diagonal composition, airborne particles and fabric motion blur, crisp subject freeze, late-afternoon window light, cinematic realism, 16:9
OpenAI: GPT-6 Luna: Model B captures the snap more vividly, with a sweeping diagonal towel, airborne flour, and three tumbling apricots, while Model A’s apricots remain on the counter and the action feels more static. Model A better matches the requested 16:9 framing, but Model B is stronger in overall motion, composition, and cinematic impact. (Second judge pass, order swapped — scores are the average of both: Model B captures the towel’s dramatic snap, airborne flour, and tumbling apricots with a vivid, cinematic | Anthropic: Claude Fable 5.1: Model B captures the dynamic whip far better —…
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 A shows exactly seven clearly separated, identical cups in a straight row and matches the requested 16:9 format. Model B also has seven countable cups, but its square format and low placement of the row are less faithful to the prompt. (Second judge pass, order swapped — scores are the average of both: Both images show exactly seven clearly countable cups in a straight row on white marble, but Model A better matches the requested 16:9 framing. Model B is clean and evenly lit, though its sq | Anthropic: Claude Fable 5.1: Both Model A and Model B render exactly seven c…
Rainy Citrus Poster
A clean stylized illustration in the manner of a vintage travel poster showing a solitary glass citrus juicer on a windowsill at dusk during rain, half a blood orange resting beside it, mood wistful and elegant; simplified geometric shapes, limited palette of teal, cream, rust, and charcoal, bold outlines, flat vector shading, refined negative space, centered composition
OpenAI: GPT-6 Luna: Model B creates a more elegant, wistful vintage travel-poster scene, with a clear rainy dusk view and a beautifully rendered glass juicer beside a half blood orange. Model A follows the simplified vector style well, but its extra objects and busier layout feel less refined. (Second judge pass, order swapped — scores are the average of both: Model B more fully captures the wistful vintage travel-poster mood, with a prominent glass juicer and half blood orange against a beautifully rendered rainy | Anthropic: Claude Fable 5.1: Model B nails the core subject—a convincingly gl…
Matchup powered by OpenRouter.