Head to head: Kling Image vs Ideogram V4.5 Edit
Kling Image vs Ideogram V4.5 Edit
By Ryan Merket · Published
A head-to-head across eight image prompts tests palette discipline, scene details, style, reflections and precise object interactions. The results reveal different strengths in visual fidelity and composition.
Ideogram V4.5 Edit takes the saved verdict with an aggregate score of **64.0** to Kling Image’s **53.6**, winning six tasks to two, with no ties. That’s a **clear result** on this sample—not a claim about statistical confidence or either model’s performance beyond it. Ideogram’s edge comes from repeatedly tracking the prompt’s specific visual requirements. It was stronger on the restricted four-color palette, the mangrove rescue ensemble, the candlelit moth archive, the ukiyo-e fishing scene and the lichen-covered rover boulder. Across these, it more reliably included the requested setting, objects and style cues. Kling had the better read on the sunlit, plant-free reading nook and the botanist’s tweezers-and-dragonfly interaction. Those wins point to real strengths in warmth and action legibility, but they didn’t outweigh the broader pattern of Ideogram matching the requested scene details more closely. This assessment applies only to the saved task sample. **Verdict: Ideogram V4.5 Edit wins this matchup, 64.0 to 53.6.**
Restricted palette
A flat-vector illustration of a mountain campsite at dusk using ONLY four colors — cream, burnt orange, teal, and charcoal — with no gradients and no stray off-palette colors, balanced composition, 16:9.
OpenAI: GPT-6 Luna: Model B better follows the four-color, flat-vector restriction, while Model A uses a prominent sky gradient and several off-palette accents. Model A fills the frame with a more immersive campsite, but Model B’s clean shapes and dusk setting make it the stronger overall match despite its generous empty space. (Second judge pass, order swapped — scores are the average of both: Model B closely follows the four-color, flat-vector constraint and depicts a campsite at dusk, though the focal elements f | Anthropic: Claude Fable 5.1: Model A violates the core constraint with a cle…
Mangrove rescue ensemble
A coherent wide gouache-style wildlife scene of seven fishing cats and two human estuary wardens working together on a low-tide mangrove flat after a storm: one warden steadies a blue crate, the other cuts loose a green nylon net snagged on pneumatophores, three cats prowl the mud near stranded mullet, two juveniles perch on a drift log watching a white-bellied sea eagle circle overhead, and an older cat stands on a half-submerged skiff at frame left; fiddler crabs, uprooted mangrove seedlings, and scattered oyster shells tie the space together, with balanced triangular composition, clear interactions among all subjects, humid pearl-gray afternoon light, 16:9.
OpenAI: GPT-6 Luna: Model B more closely matches the gouache-style mangrove rescue scene, with clear warden actions, a cat on the skiff, kittens on the drift log, stranded fish, and soft estuary light. Model A is vivid and detailed, but its heavy rain, less evident mangrove habitat, and fewer cats diverge more from the prompt. (Second judge pass, order swapped — scores are the average of both: Model B more closely matches the requested gouache-like, pearl-gray low-tide rescue scene, with clear warden actions and a | Anthropic: Claude Fable 5.1: Model B delivers a cleaner gouache/watercolor lo…
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 delivers the dramatic close-up and shows strong window reflections across the chrome and glass, with the apple visibly represented through the water; Model B includes all the main objects but the apple and window are not clearly reflected in the water. Model A also has the more polished, visually striking composition. (Second judge pass, order swapped — scores are the average of both: Model A more clearly shows the window and red apple reflected or refracted in both the chrome and | Anthropic: Claude Fable 5.1: Model A is far more photographic and dramatic,…
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 more clearly delivers the requested warm, sunlit nook with the armchair, book stack, and tea, while keeping the frame free of plants, lamps, and artwork. Model B is also attractive and well composed, but the scattered leaves and cooler, more muted light make it slightly less faithful to the prompt. (Second judge pass, order swapped — scores are the average of both: Both images include the requested reading-nook elements and avoid plants, lamps, and artwork; Model A feels especially cozy, | Anthropic: Claude Fable 5.1: Both outputs fully respect the negation constra…
Candlelit moth archive
A moody cinematic interior painting of a nocturnal natural history archive lit only by three beeswax candles and a single oil lamp, where a drawer of enormous moon moth specimens lies open on a walnut table beside handwritten labels, glass vials of pollen, and a brass magnifier; the warm candlelight creates soft falloff, deep amber shadows, and glowing translucent wings while cool blue rain leaks through a high attic window, setting a precise midnight atmosphere, composed from a low three-quarter angle with rich chiaroscuro and intimate stillness, 16:9.
OpenAI: GPT-6 Luna: Model B more clearly reads as a natural-history archive and a moody interior painting, with oversized moon moths, the rainy attic window, and the requested candle-and-lamp arrangement. Model A has especially strong chiaroscuro and a focused composition, but feels more photorealistic and its moth specimens are less imposing. (Second judge pass, order swapped — scores are the average of both: Model B better captures the enormous, glowing moon moths and the painterly intimacy of a candlelit archive | Anthropic: Claude Fable 5.1: Model A nails the moody midnight chiaroscuro, s…
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 depicts a fishing boat riding a dramatic cresting wave at dawn, with fishermen and a rising sun reinforcing the scene while retaining ukiyo-e linework and stylized foam. Model A has a strong woodblock-print look, but its boat is less legible and the dawn setting is less explicit. (Second judge pass, order swapped — scores are the average of both: Model B more clearly depicts fishermen in a boat riding a cresting wave at dawn, with the sun, muted indigo-and-cream palette, bol | Anthropic: Claude Fable 5.1: Model A captures authentic woodblock texture an…
Bog botanist hands
A field-journal style realistic wildlife portrait of a bog botanist kneeling on a cedar plank over a red sphagnum marsh, clearly visible from head to knees, holding a carnivorous pitcher plant in her left hand while her right hand uses fine brass tweezers to remove a trapped dragonfly wing; both hands are fully shown with all five fingers in natural positions, accurate wrist angles, and believable body proportions, her rain-freckled face intent behind round glasses, waxed olive jacket, mud-stained trousers, and a canvas satchel of specimen tags, with sundew and cotton grass around her, soft overcast morning light, shallow depth of field, 16:9.
OpenAI: GPT-6 Luna: Model A more clearly depicts the key action of using tweezers at a dragonfly beside a pitcher plant, though the botanist’s head is cropped and the hand anatomy is imperfect. Model B shows the full kneeling figure, cedar plank, and specimen satchel more completely, but its dragonfly is separate from the tweezers and the requested interaction is unclear. (Second judge pass, order swapped — scores are the average of both: Model A makes the botanist’s hands, pitcher plant, tweezers, and dragonfly in | Anthropic: Claude Fable 5.1: Model B shows the botanist fully from head to k…
Lichen rover texture
An ultra-detailed isometric 3D render of a palm-sized alpine survey rover parked beside a glacial boulder covered in silver-green map lichen and black crustose speckles, every surface rendered crisply: braided cable sleeves, knurled titanium screws, machined wheel treads packed with powdery scree, frost crystals on a brushed aluminum chassis, frayed orange safety ribbon, a folded ripstop sample bag, and tiny cushions of moss between quartz grains, lit by cold high-altitude daylight with razor-sharp focus and no smudging, 16:9.
OpenAI: GPT-6 Luna: Model B more clearly depicts the silver-green lichen and black speckles, along with the cable sleeves, orange ribbon, sample bag, and rocky scree. Model A has a cleaner isometric presentation and sharper overall focus, but its boulder reads mainly as icy rather than lichen-covered. (Second judge pass, order swapped — scores are the average of both: Model B more directly captures the silver-green lichen and black speckles on the glacial boulder, along with scree, cable sleeves, and the orange saf | Anthropic: Claude Fable 5.1: Model B nails the silver-green map lichen with…
Matchup powered by OpenRouter.