Head to head: Ideogram V4.5 Edit vs Kling Image
Ideogram V4.5 Edit vs Kling Image
By Ryan Merket · Published
A practical image-model face-off across anatomy, composition, text, and tightly constrained briefs. The results turn on more than polish: framing and prompt adherence repeatedly decide which image is actually usable.
Kling Image takes this matchup, 58.8 to 51.4, with a tentative lean and six task wins to Ideogram V4.5 Edit’s two. That’s a clear result, not a photo finish: Kling more often delivered the requested composition instead of a polished image in the wrong shape. The pattern is especially clear in the 16:9 briefs. Kling framed the candid handwork scene correctly, built out the ferry diagram with its kiosk and route board, and gave the rainy taxi rank three receding cars and a readable “Velmor Central” sign. Its more natural hands also helped in the mechanic scene, where Ideogram’s ticket punch was clearer but its oversized foreground hands overwhelmed the portrait. Ideogram’s wins show where it still has an edge: it rendered the depot poster text more cleanly, and its bedroom layout was easier to read. But the square framing cost it repeatedly in tasks that explicitly asked for widescreen. Kling wasn’t flawless, either: both models strayed from the locked palettes, and its poster included a duplicated route tag and a blemish in “BAY.” **Verdict: Kling Image is the stronger all-rounder here. It wins by following the brief more consistently—especially on framing, object relationships, and believable anatomy—while Ideogram’s sharper text and cleaner spatial presentation are real but narrower advantages.**
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 B better matches the requested 16:9 candid scene, with both hands clearly visible and a plausible grip on the colorful threads. Model A has appealing soft daylight and shallow depth of field, but its square framing misses the requested aspect ratio and the hand positions look slightly less natural. (Second judge pass, order swapped — scores are the average of both: Model B more closely matches the requested 16:9 candid photo, with both hands clearly visible and distinct colorful threads be | Anthropic: Claude Fable 5.1: Model A has pleasant natural outdoor light and…
Palette-Locked Ferry Diagram
Create a clean isometric 3D render of a compact electric river ferry docked beside a tiny ticket kiosk and boarding ramp, shown in a tidy harbor cutaway composition, 16:9. Restrict the entire image to this exact named color palette only: Japanese traditional palette — Shironeri (#FCFAF2), Sumi (#1C1C1C), Ruri-iro (#1F4788), Hiwa-iro (#BEC23F), Karakurenai (#D7003A), and Kincha (#C66B27). No gradients outside these hues, no stray off-palette colors, no realistic sky blues or skin tones beyond the listed palette. The ferry should have geometric windows, stacked life rings, mooring posts, and a route map board, rendered as crisp flat-shaded forms with strong edge definition and controlled shadows.
OpenAI: GPT-6 Luna: Model B better matches the requested 16:9 harbor cutaway and clearly includes the kiosk, route map board, and ferry, while Model A has a strong ferry-and-ramp arrangement but is square. Both substantially violate the palette lock with off-palette water, surfaces, and shading; Model B’s layout and object coverage give it the edge. (Second judge pass, order swapped — scores are the average of both: Model B better matches the requested 16:9 harbor cutaway and includes the kiosk, route map, ramp, li | Anthropic: Claude Fable 5.1: Model B delivers the requested 16:9 harbor cuta…
Mechanic With Ticket Punch
Show a full upper-body portrait of a transit mechanic standing beside an open vintage trolley doorway, clearly visible from head to mid-thigh, holding a brass ticket punch in one hand and a folded inspection card in the other, with both hands fully in frame and anatomically correct. She wears a navy work jacket with reflective piping, charcoal coveralls, and a pale green neck scarf; her pose is relaxed but natural, one shoulder leaning lightly against the vehicle, fingers separated and believable, nails subtle, wrists and elbows aligned correctly, and body proportions realistic. Render in a polished editorial illustration style with fine detail, soft overcast depot lighting, shallow depth of field, and careful emphasis on accurate hands, knuckles, and facial anatomy.
OpenAI: GPT-6 Luna: Model A more clearly depicts a ticket punch and folded inspection card, but its oversized foreground hands feel distorted and dominate the portrait. Model B has a more natural pose, convincing hands, and a stronger composition, though its brass tool is less clearly a ticket punch and the card is small. (Second judge pass, order swapped — scores are the average of both: Model B has a relaxed pose, clearly visible hands with believable anatomy, and soft depot lighting, while Model A’s extreme fore | Anthropic: Claude Fable 5.1: Model A does depict a recognizable brass ticket…
Rainy Taxi Rank Realism
Create a photorealistic evening street scene of a taxi rank outside the fictional "Velmor Central" station just after rain, 16:9, viewed from curb height at a slight angle so three vehicles recede into the distance. A brushed aluminum luggage cart, wet oak bench slats, yellow painted curb markings, a driver in a damp wool coat with visible skin texture, and a black sedan with realistic clearcoat reflections must all look physically convincing. Use mixed lighting from sodium-vapor streetlamps, cool station LEDs, and soft interior taxi dome lights, producing believable color separation, reflections in puddles, contact shadows under tires, water beading on metal, and accurate material response on glass, chrome, fabric, wood, and skin.
OpenAI: GPT-6 Luna: Model B better matches the requested 16:9 curbside scene, with three receding vehicles, a convincing wet black sedan, warm and cool mixed lighting, and a readable “Velmor Central” sign. Model A has strong wet-surface detail and a prominent luggage cart, but misses the requested aspect ratio and its station text is unclear. (Second judge pass, order swapped — scores are the average of both: Model B more closely matches the requested evening taxi rank, with three receding vehicles, a readable Velm | Anthropic: Claude Fable 5.1: Model B delivers the 16:9 curb-height framing,…
Depot Poster Text Test
Design a retro-modern screenprinted safety poster mounted on a tram depot wall, photographed straight-on for maximum legibility, 16:9. The poster must contain exactly this text, clearly readable and correctly spelled, with strong typographic hierarchy: "NORTH QUAY TRAMWORKS" on the top line, "BAY 7" large in the center, and "HELMETS ON AFTER 21:40" below. Include a small route tag reading "RQ-17" in one corner. The scene should show the poster attached to corrugated metal beside a parked amber maintenance tram, but the text remains the focal point; use crisp sans-serif lettering, high contrast ink on paper, realistic print texture, and even industrial lighting with no distortion, gibberish, or extra words.
OpenAI: GPT-6 Luna: Model B better matches the requested 16:9, straight-on depot-poster composition and keeps the required wording highly legible, though its duplicated route tag and blemish within “BAY” detract from accuracy. Model A has cleaner poster text and a polished look, but its square framing misses the specified aspect ratio and its top heading wraps onto two lines. (Second judge pass, order swapped — scores are the average of both: Model B better matches the requested 16:9 layout, single-line heading, an | Anthropic: Claude Fable 5.1: Model A renders all required poster text cleanl…
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: Model B more closely follows the requested 16:9 tabletop composition, with the duck atop the green cube and the blue cylinder behind the objects. Model A also places the objects correctly, but its square framing and oversized cylinder diverge more from the prompt. (Second judge pass, order swapped — scores are the average of both: Model B more closely matches the photorealistic 16:9 tabletop scene, with the requested colors and object relationships clearly depicted. Model A includes the same obj | Anthropic: Claude Fable 5.1: Model B satisfies every constraint — correct co…
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 A conveys dusk more clearly and has a richer, well-balanced campsite scene, but its square format misses the requested 16:9 ratio. Model B matches the ratio and includes recognizable campsite details, though its dusk atmosphere is weaker and both images use noticeable color variations beyond a strict four-color palette. (Second judge pass, order swapped — scores are the average of both: Model B better follows the flat-vector 16:9 brief and keeps the campsite prominent, though it includes s | Anthropic: Claude Fable 5.1: Model A is the more polished and charming illus…
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 A follows the requested arrangement clearly, with the bed along the left wall, a round rug near the floor center, a desk beneath the back-wall window, and a lamp in the front-right corner. Model B includes the same key elements, but its stark, simplified rendering and less clearly defined window and room details make it less polished and readable. (Second judge pass, order swapped — scores are the average of both: Both models depict the requested isometric bedroom layout, but Model A has a | Anthropic: Claude Fable 5.1: Model A places the bed on the left wall, the de…
Matchup powered by OpenRouter.