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
Kling Image and Google’s Nano Banana Pro (Gemini 3 Pro Image) face off across eight image-generation tasks, from exact text layout to richly specified scenes. The comparison turns on how well each model follows visual constraints while keeping the resulting image coherent and legible.
Google: Nano Banana Pro (Gemini 3 Pro Image) takes the saved task sample **7–1, with no ties**, and leads on aggregate score, **65.1 to Kling Image’s 57.4**. That makes this a decisive editorial result for this sample—not a claim of statistical certainty or a verdict on every kind of image generation. The gap shows up most in scenes where several details must work together. Nano Banana Pro better conveyed the rookery’s coastal setting, placed the bedroom elements coherently, kept the aurora harbor boat-free, and depicted the juvenile swan’s stretcher transfer more clearly. It also handled the requested dawn mood and reflective chrome teapot more convincingly. Kling’s win was specific and meaningful: its three-line text design kept the requested wording crisp, correctly spelled, and centered. Nano Banana Pro rendered the words accurately too, but broke the requested layout and turned it into a poster mockup—a reminder that polished output is not the same as following the brief. **Verdict: Nano Banana Pro is the stronger pick in this saved task sample, especially for detailed, constraint-heavy scenes; Kling Image’s standout advantage here is precise text layout.** This assessment applies only to the saved task sample.
Rookery permit sign
A clean documentary-style photograph of a weathered park sign at the entrance to a seabird rookery, mounted on driftwood posts among dune grass, with highly legible dark navy sans-serif text on a matte pale cream background that reads exactly: "Tern Cay Preserve\nZONE B\nKeep 40 m Back"; bright overcast daylight for even readability, straight-on composition, sharp focus, 16:9.
OpenAI: GPT-6 Luna: Model B better conveys the seabird rookery with birds and coastal dune context while keeping the requested wording highly legible. Model A is exceptionally clear and straight-on, but its tighter framing provides less of the rookery setting. (Second judge pass, order swapped — scores are the average of both: Model B more fully matches the rookery setting, with seabirds, shoreline, dune grass, weathered driftwood posts, and the exact, legible sign text. Model A is also sharp and well composed, but | Anthropic: Claude Fable 5.1: Both render the three-line text exactly and leg…
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 has a particularly cohesive ukiyo-e composition, while retaining the bold indigo waves and stylized foam. Model A is also strongly on-style and closer to the muted palette, but its composition feels less balanced and the dawn setting is less explicit. (Second judge pass, order swapped — scores are the average of both: Model B better fulfills the dawn setting with a visible rising sun and a muted indigo-and-cream palette, while both entrie | Anthropic: Claude Fable 5.1: Both lean heavily on Hokusai's Great Wave and s…
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 B more convincingly renders the chrome teapot and clearly shows the window and red apple reflected in its surface, with strong side lighting. Model A has a pleasing scene and puts the apple on the counter, but its teapot reads as brass rather than chrome and the required reflections are less clear, especially in the water. (Second judge pass, order swapped — scores are the average of both: Model B more closely follows the prompt with a clearly chrome teapot, dramatic side light, and visibl | Anthropic: Claude Fable 5.1: Model B delivers a genuinely mirror-like chrome…
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 bed, desk beneath the back-wall window, centered round rug, and front-right lamp more clearly in a cohesive, polished isometric composition. Model A includes the requested elements but feels sparser, and its desk and rug placement are less convincing. (Second judge pass, order swapped — scores are the average of both: Model B clearly places the bed along the left wall, the desk beneath the back-wall window, the round rug near the center, and the lamp in the front-right corner, | Anthropic: Claude Fable 5.1: Model B places the bed flush against the left w…
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 follows the requested three-line layout, with all text crisp, correctly spelled, and centered on a deep navy textured background. Model B renders the text accurately too, but splits the headline across two lines and presents the design as a poster mockup, departing from the requested layout. (Second judge pass, order swapped — scores are the average of both: Model A presents the exact three requested lines with legible, correctly spelled text and strong centering. Model B renders the wor | Anthropic: Claude Fable 5.1: Model A renders all three lines exactly as spec…
Empty aurora harbor
A remote Arctic fishing harbor at night under vivid green aurora, with wooden docks, stacked crab pots, sodium-vapor lamps, frost on ropes, and two bundled naturalists watching the sky from shore — but absolutely no boats anywhere in the water or tied to the docks; crisp photorealism, wide 16:9 composition, still black water reflecting the lights, quiet wintry atmosphere.
OpenAI: GPT-6 Luna: Model A has a striking aurora and harbor, but visibly includes a boat tied to the dock, directly violating a key constraint. Model B keeps the water boat-free and includes the bundled observers, docks, crab pots, lamps, and vivid aurora, though its contrast and detail feel somewhat overprocessed. (Second judge pass, order swapped — scores are the average of both: Model B captures the empty, wintry harbor with two observers, stacked crab pots, frosted docks, warm lamps, and vivid aurora reflected | Anthropic: Claude Fable 5.1: Model A directly violates the key negative cons…
Wetland rescue tableau
At blue-hour in a misty coastal marsh, a wildlife rescue team in olive rain gear carefully transfers a juvenile whooper swan from a canvas stretcher into a low aluminum skiff while two border collies wait on the reeds, a biologist kneels beside open waterproof cases of banding tools, and three black-necked stilts lift off behind them; cinematic realism, coherent multi-subject scene composition, balanced triangular arrangement across the frame, rippling reflections, cool dawn light with one warm lantern glow, 16:9.
OpenAI: GPT-6 Luna: Model B more clearly depicts the careful stretcher transfer of a gray juvenile swan, with the rescue team, dogs, equipment, birds in flight, lantern, and reflective marsh all legible in a coherent scene. Model A has a pleasingly wide arrangement and strong reflections, but the swan reads as an adult and the transfer is less convincing. (Second judge pass, order swapped — scores are the average of both: Model B more clearly depicts the juvenile gray swan being carried on a canvas stretcher into t | Anthropic: Claude Fable 5.1: Model B renders an actual transfer of a correct…
Snow leopard remorse
A close portrait of a snow leopard perched on a lichen-covered rock, convincingly expressing remorse after knocking over a small field researcher’s tin lunch pail visible near one paw; hyper-detailed painterly wildlife illustration, ears slightly pinned back, brows softened, eyes glossy and downward, mouth subtly tense, whiskers damp with melting snow, pale overcast mountain light, shallow depth of field isolating the face.
OpenAI: GPT-6 Luna: Model B better captures the requested close portrait and remorseful downward gaze, with the open lunch pail, lichen-covered rock, and painterly mountain setting all clearly visible. Model A has a convincing snow leopard and rock, but the wider framing and less expressive face make it a weaker match. (Second judge pass, order swapped — scores are the average of both: Model B better matches the requested close portrait, with a lichen-covered rock, spilled lunch pail beside the paw, and a softly do | Anthropic: Claude Fable 5.1: Model B delivers the requested close portrait w…
Matchup powered by OpenRouter.