Head to head: Kling Image vs Google: Nano Banana Pro (Gemini 3 Pro Image)
Kling Image vs Google: Nano Banana Pro (Gemini 3 Pro Image)
By Ryan Merket · Published
Eight varied image prompts put instruction-following, composition, style control, and detail legibility to the test. The results show where each model is reliable—and where a polished image can still miss the brief.
Nano Banana Pro takes this matchup on the work itself: it wins six of eight tasks, with one tie, and leads the aggregate score 64.8 to 58.6. The statistical verdict gives Google a 96% win confidence, a decisive result—not a claim that every image was better. Its advantage is strongest when the prompt has several constraints to satisfy at once. Nano Banana more faithfully handled the colored-object relationships, omitted forbidden items in the reading nook, and delivered a more convincing ukiyo-e scene. It also made the “L-47” tag readable on the terrarium and kept the breakfast illustration sparse, as requested. In the laundry scene, its receding arrangement and corner-angle perspective were clearer, even if the warm lighting and three button jars weren’t exact. Kling’s win came in the sunset teapot cutaway, where the centered scene and visible rosemary and apricots read more cleanly. Nano Banana’s poster-like treatment muddied the composition with overlapping teapot shapes and added text. The hands task was a tie: Kling made the grip easier to inspect, while Nano Banana offered more context, but neither separated itself overall. **Final call: Google’s Nano Banana Pro is the stronger choice for prompt fidelity across varied, detail-heavy scenes. Kling has a clear win when a clean, legible composition matters, but it didn’t match Nano Banana’s consistency across the full set.**
Attribute binding
A photorealistic tabletop scene: a matte green cube to the LEFT of a glossy red sphere, a small blue cylinder BEHIND both, and a single yellow rubber duck sitting ON TOP of the cube. Colors must stay on the correct objects. Soft studio light, 16:9.
OpenAI: GPT-6 Luna: Both models place the yellow duck on the green cube, with the red sphere to the right and blue cylinder behind; Model B better captures the duck as small and the cube as matte, with a more balanced tabletop composition. Model A is also clear and photorealistic, but its harsher shadows and oversized duck make it slightly less faithful to the prompt. (Second judge pass, order swapped — scores are the average of both: Model B follows the requested object colors and spatial relationships, with the d | Anthropic: Claude Fable 5.1: Both models bind every color to the correct obj…
Negation
A cozy reading nook with an armchair, a stack of books, and a mug of tea by a window — with absolutely NO plants, NO lamps, and NO artwork or picture frames anywhere in the frame. Warm afternoon light, 16:9.
OpenAI: GPT-6 Luna: Model A captures the chair, books, tea, and warm window light, while Model B presents them in a more complete and inviting reading nook with the books and mug beside the armchair. Both avoid plants, lamps, and artwork, but Model B’s broader, balanced composition feels more naturally cozy. (Second judge pass, order swapped — scores are the average of both: Model B creates a more complete, balanced reading nook with the armchair, books, and mug arranged beside the window; Model A includes the requ | Anthropic: Claude Fable 5.1: Both Model A and Model B correctly omit plants,…
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 fully evokes a ukiyo-e print, with stylized foam, a muted indigo-and-cream palette, and a dawn sun that strengthens the scene. Model A has a striking cresting wave and prominent boat, but the boat’s oversized, bare rigging makes the composition feel less cohesive and the dawn setting less clear. (Second judge pass, order swapped — scores are the average of both: Model B more fully captures the ukiyo-e tradition with a fishing boat and crew riding the wave, crisp stylized foam, and a | Anthropic: Claude Fable 5.1: Model B delivers a more convincing ukiyo-e prin…
Atrium Terrarium Refractions
A hyper-real still life of a hexagonal glass terrarium on a smoked-glass dining table, containing black pebbles, one striped calathea leaf, and a tiny brass key tagged "L-47"; behind it, a rain-streaked apartment atrium window and a chrome desk lamp must appear correctly reflected and refracted through every pane, with believable distortions, internal reflections, and soft caustics on the tabletop, cool overcast morning light, eye-level composition, 16:9.
OpenAI: GPT-6 Luna: Model B more clearly depicts the hexagonal paneled terrarium and renders the “L-47” tag legibly, while also including the rainy atrium, chrome lamp, and strong tabletop reflections. Model A has appealing glass detail and atmosphere, but the key lacks its requested tag and the terrarium reads less distinctly as hexagonal. (Second judge pass, order swapped — scores are the average of both: Model B more fully matches the requested angular terrarium and includes a clearly readable “L-47” tag, while | Anthropic: Claude Fable 5.1: Model B renders a true eye-level composition wit…
Breakfast Table Without Cutlery
A quiet early-morning kitchen scene showing a round oak breakfast table with a blue bowl of figs, a half-sliced rye loaf on a wooden board, a small jar of quince jam, a ceramic mug with steam, and a linen napkin, rendered in minimalist line art with sparse flat color accents and gentle window light—deliberately no cutlery anywhere: no knife, no spoon, no fork, and no metal utensils visible in the room.
OpenAI: GPT-6 Luna: Model B better matches the requested minimalist line art with sparse color accents, while both images include the breakfast items and avoid visible cutlery. Model A is more richly rendered and has a detailed kitchen background, but that makes it busier and less faithful to the specified style. (Second judge pass, order swapped — scores are the average of both: Model B more closely matches the requested minimalist line art with sparse color and a quiet, uncluttered breakfast scene, while includin | Anthropic: Claude Fable 5.1: Model B correctly delivers the requested minima…
Sunset Teapot Cutaway
A clean stylized vector illustration of a dented enamel teapot with a missing lid, revealing three sprigs of rosemary and two apricots inside, sitting on a narrow windowsill during a dusty orange sunset; crisp geometric shapes, minimal shading, elegant mid-century poster design, muted terracotta and teal palette, calm slightly wistful mood, strong side light casting long simple shadows, centered composition, 16:9.
OpenAI: GPT-6 Luna: Model A more clearly delivers the centered sunset windowsill scene, with the rosemary and two apricots visible inside a crisp teal teapot; the raised lid is a small departure from the requested missing lid. Model B has appealing poster styling, but the oversized overlapping teapot shapes and added text make the composition less clear and less faithful to the prompt. (Second judge pass, order swapped — scores are the average of both: Model A has a clearer centered windowsill scene, stronger sunse | Anthropic: Claude Fable 5.1: Model A delivers a clean, centered teal enamel…
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 gives a clearer close view of both hands and their plausible pinching grip on the bracelet threads, with soft daylight and shallow depth of field. Model B has a more contextual candid setting, but the hands and thread manipulation are less clearly legible. (Second judge pass, order swapped — scores are the average of both: Model A makes the hands and their grip on the bracelet threads clearer, with convincing anatomy and crisp detail against a softly blurred outdoor background. Model B h | Anthropic: Claude Fable 5.1: Model A delivers a tight, legible view of both…
Laundry Shelf Vanishing Point
A meticulously composed domestic still life of a utility-room shelving wall seen from a low corner angle, with one clear vanishing point: a tall bottle of cedar detergent in the foreground, then progressively smaller objects receding along the shelves—a folded sunflower-yellow towel stack, four labeled spice jars repurposed for buttons, a mint toolbox, a hand vacuum, and a tiny radio—every object sized consistently by distance, straight shelf lines, no warped geometry, bright neutral daylight, photographic realism, 16:9.
OpenAI: GPT-6 Luna: Model B presents the requested objects in a more legible receding sequence with a strong corner-angle vanishing point, while Model A separates several objects across shelves and has less coherent perspective. Model A is brighter and more neutral, but Model B’s arrangement and readable detergent label better support the prompt despite its warmer light and only three button jars. (Second judge pass, order swapped — scores are the average of both: Model B follows the requested low corner-angle sequ | Anthropic: Claude Fable 5.1: Model B delivers a convincing single-vanishing-…
Matchup powered by OpenRouter.