Head to head: Wan vs Google: Nano Banana Pro (Gemini 3 Pro Image)
Wan vs Google: Nano Banana Pro (Gemini 3 Pro Image)
By Ryan Merket · Published
We compared Wan and Google’s Nano Banana Pro across eight image-generation tasks, from precise room layouts and reflective surfaces to named art styles and intricate still lifes. The results test both prompt fidelity and how well each model turns a brief into a coherent image.
Google takes the matchup with **65.3 to Wan’s 58.3**, winning six tasks to Wan’s one, with one tie. The statistical verdict gives Google a **limited confidence** win: a substantial overall advantage, not a coin flip dressed up as a result. That edge came from breadth. Google won on spatial layout, the Velis Night Courier scene, Orbital Relief Cry, hands and anatomy, named art style, and Clockwork Moss Reliquary. Across those varied briefs, it more often delivered the requested composition and mood together—from a desk placed beneath the window to a readable, dramatic ukiyo-e dawn scene. Wan’s clearest win was reflections and glass. Its image included the actual apple and made the chrome and countertop reflections convincing, while Google’s reflected apple was not clearly present as a physical object. The Kestrel Quay Notice was a tie: both models rendered the exact wording legibly, with Google leaning into the foggy starport atmosphere and Wan offering a rain-damp industrial setting. **Final call: Google’s Nano Banana Pro is the stronger model in this head-to-head, winning decisively across the task set; Wan’s standout reflection work is real, but not enough to close the gap.**
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 B places the desk directly beneath the window and keeps the rug centered and lamp in the front-right corner, matching the requested layout more closely. Model A has a crisp flat-vector style, but its desk sits beside rather than beneath the window. (Second judge pass, order swapped — scores are the average of both: Model B places the bed against the left wall, the desk beneath the window, the rug near the floor’s center, and the lamp in the front-right corner. Model A is clean and well com | Anthropic: Claude Fable 5.1: Model B places the desk directly under the wind…
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 presents all requested objects in a polished, photorealistic composition, with convincing chrome and countertop reflections. Model B shows the window and apple clearly reflected in the teapot, but the actual apple is mostly absent from the scene and its appearance in the glass reads more like an object seen through the water than a clear surface reflection. (Second judge pass, order swapped — scores are the average of both: Model A includes the apple physically on the counter and shows a | Anthropic: Claude Fable 5.1: Model A delivers a coherent, photorealistic sce…
Velis Night Courier
A clean stylized illustration of a lone night courier on a hovering scooter crossing the glass causeway of Velis-9 under light neon rain, wearing a long angular cloak and carrying a glowing parcel cube; minimalist line art with flat vector shapes, limited palette of indigo, mint, and coral, elegant sci-fi city silhouettes fading into mist, calm but slightly wistful mood, strong negative space, side view composition, 16:9.
OpenAI: GPT-6 Luna: Model B more clearly depicts the courier riding in side view, with a crisp angular cloak, glowing parcel, rain, and a glass causeway in a cohesive minimalist palette. Model A conveys hovering more explicitly, but the courier’s standing pose and busier detail make the scene less aligned with the intended composition. (Second judge pass, order swapped — scores are the average of both: Model B more clearly depicts a courier riding a scooter in side view, with the angular cloak, glowing coral parcel | Anthropic: Claude Fable 5.1: Model A better captures the brief's misty fadin…
Orbital Relief Cry
A close-up cinematic portrait of a young asteroid-rescue pilot in a scuffed pressure suit inside a dim salvage shuttle, visibly overwhelmed with relief after spotting a survivor beacon—eyes glossy with tears, mouth half-laughing, half-crying, cheeks tense, brows lifted in disbelief; cool cyan instrument glow on one side of the face, warm amber emergency light on the other, shallow depth of field, highly realistic sci-fi photography, intimate framing, 16:9.
OpenAI: GPT-6 Luna: Model A more completely captures the prompt with the scuffed pressure helmet, unmistakable tearful relief, intimate framing, and strong cyan–amber split lighting. Model B is also cinematic and emotionally convincing, but its expression is subtler and the visible suit lettering is distracting. (Second judge pass, order swapped — scores are the average of both: Model B more clearly conveys overwhelmed relief through the tearful, half-laughing expression, and its scuffed suit, shuttle instruments, | Anthropic: Claude Fable 5.1: Model B delivers a more convincing half-laughing…
Kestrel Quay Notice
A rain-damp metal notice board at a foggy starport cargo pier at dawn, photographed straight-on for maximum readability, featuring a clean printed transit placard with large legible text that reads exactly: "GATE 7B\nKESTREL QUAY\nAUTHORIZED CREW ONLY"; bold sans-serif typography, crisp high-contrast ink on pale composite panel, small rivets and water droplets around the sign, realistic industrial sci-fi setting, 16:9.
OpenAI: GPT-6 Luna: Model A renders the exact text clearly and has a convincing rain-damp industrial setting, while Model B pairs equally legible text with a stronger foggy starport atmosphere and more polished composition. Model B’s sign is slightly weathered rather than clean, but it remains highly readable and fits the prompt well. (Second judge pass, order swapped — scores are the average of both: Model B better captures the foggy, rain-damp starport setting and presents the exact placard text prominently and l | Anthropic: Claude Fable 5.1: Both render the three-line text flawlessly in b…
Hands & anatomy
A natural candid photo of a person tying a friendship bracelet, both hands clearly visible with anatomically correct fingers and a plausible grip on the threads, soft daylight, shallow depth of field, 16:9.
OpenAI: GPT-6 Luna: Model A clearly shows both hands manipulating the bracelet threads, with more legible, natural finger positions and a convincing close-up of the craft. Model B has a pleasing candid scene and soft daylight, but the wrist-tying action obscures some of the hand anatomy and thread grip. (Second judge pass, order swapped — scores are the average of both: Model A more clearly shows both hands working the threads with a plausible grip and anatomically convincing fingers, while Model B has a warm candi | Anthropic: Claude Fable 5.1: Model A renders crisp, anatomically sound hands…
Named art style
A ukiyo-e woodblock print of a fishing boat riding a large cresting wave at dawn, faithful to the flat color planes, bold outlines, and stylized foam of the tradition, muted indigo and cream palette.
OpenAI: GPT-6 Luna: Model B more clearly conveys dawn with the rising sun and places visible fishing boats within a dramatic, well-balanced ukiyo-e wave scene. Model A has a strong muted palette and classic woodblock look, but its boats are less legible and the dawn setting is less explicit. (Second judge pass, order swapped — scores are the average of both: Model A more consistently captures the flat color planes, bold outlines, and stylized foam of ukiyo-e, with a clear fishing boat riding the cresting wave. Mode | Anthropic: Claude Fable 5.1: Model A adheres more tightly to the muted indig…
Clockwork Moss Reliquary
An ultra-detailed macro still life of a forgotten brass reliquary shaped like a many-ringed gyroscope, half-swallowed by velvety emerald moss and silver-thread lichens on a stone shelf in a ruined observatory; every surface rendered sharply—fine engraved numerals, tarnish gradients, tiny screws, frayed silk wrapping, granular dust, delicate fern spores, and dense moss fibers—lit by a narrow shaft of cold morning light from above, dark background, high-end photorealism.
OpenAI: GPT-6 Luna: Model A more clearly delivers the many-ringed, engraved brass reliquary, frayed wrapping, and sharply rendered surface details, while Model B has especially evocative light and atmosphere but a softer, less legible central object. Model A’s centered still-life framing also better suits the requested macro detail. (Second judge pass, order swapped — scores are the average of both: Model B feels more convincingly forgotten and half-swallowed by moss, with atmospheric cold light and richly textured | Anthropic: Claude Fable 5.1: Model B delivers a convincing macro still life…
Matchup powered by OpenRouter.