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-generation tasks, from exact object counts and room layouts to illustrated posters and named art styles. The comparison tests both prompt compliance and the visual choices each model makes along the way.
Ideogram V4.5 Edit takes the verdict: **62.5** to Kling Image’s **57.5**, with five task wins to Kling’s two and one tie. This is a clear editorial result on the saved task sample—not a claim about statistical significance or performance beyond it. Ideogram’s edge came across varied briefs: it won the parking-meter blueprint, ramen poster, restricted-palette scene, repair-block-party image, and ukiyo-e fishing-boat task. Its strongest results tended to combine the requested scene details with a coherent, finished composition; the ramen poster’s correct bag text and rain-slick Osaka setting are a particularly concrete example. Kling had clear wins of its own. Its seven-cup image was more uniform and easier to count, and its bedroom followed the requested isometric layout more faithfully. The brakeflip prompt ended in a tie: each model made a different part of the action brief work, with neither taking the overall task verdict. The saved outcomes favor Ideogram, but the margin belongs to this particular mix of prompts. **Verdict: Ideogram V4.5 Edit wins this head-to-head, 62.5–57.5.**
Nine Parking Meters Blueprint
A precise blueprint-style technical illustration of exactly nine distinct vintage parking meters lined along a Brooklyn curb for a municipal restoration catalog, each meter fully visible and individually countable, no duplicates overlapping, each with a different dent pattern or sticker, drawn in white and pale cyan drafting lines on deep navy with faint measurement annotations, orthographic street-elevation composition with even studio-like lighting.
OpenAI: GPT-6 Luna: Model A more closely follows the requested orthographic curbside elevation, with exactly nine fully visible meters and a clear curb; Model B has more refined drafting details but uses a perspective ground grid and has inconsistent meter labels. Model A’s final two meters face backward and several meters are less individually distinctive, while Model B’s linework and catalog presentation are especially polished. (Second judge pass, order swapped — scores are the average of both: Model B better ca | Anthropic: Claude Fable 5.1: Model B delivers a genuine blueprint plate with…
Brakeflip Over Tram Tracks
A street trials rider captured mid-action performing a brakeflip over wet tram tracks on a steep Lisbon avenue, front wheel twisted sideways, coat tails snapping, loose receipts and orange leaves whipping through the air, dynamic low-angle composition with strong motion blur in the background but sharp focus on the rider, dramatic late-afternoon sun cutting between apartment facades, high-energy editorial sports photography, 16:9.
OpenAI: GPT-6 Luna: Model B more clearly establishes the steep Lisbon avenue, wet tram rails, and tram, with motion blur and a snapping coat adding energy to the stunt. Model A has striking late-afternoon light and a strong close-up action pose, but the tracks and brakeflip are less clearly conveyed. (Second judge pass, order swapped — scores are the average of both: Model A makes the brakeflip-like maneuver the clear focal point and pairs it with striking low-angle sunlight, wet tracks, receipts, and orange leaves | Anthropic: Claude Fable 5.1: Model A delivers a punchy, sharply lit editoria…
Exact counting
A clean overhead flat lay on white marble of EXACTLY seven identical espresso cups arranged in a single straight row, each clearly separate and individually countable, even soft lighting, 16:9.
OpenAI: GPT-6 Luna: Model A shows exactly seven clearly countable, matching espresso cups in a single straight row on white marble. Model B also has seven cups, but the added saucers and spoons and the smaller, less prominent cups make it less faithful to the requested clean flat lay. (Second judge pass, order swapped — scores are the average of both: Both images show exactly seven cups in a straight row, but Model A presents them more uniformly and clearly as identical espresso cups. Model B includes a spoon and m | Anthropic: Claude Fable 5.1: Model A delivers exactly seven clearly identica…
Neon Ramen Cyclist Poster
A clean vector illustration of a lone bicycle courier stopped under a flickering ramen shop sign on a rain-slick side street in Osaka, wearing a mustard yellow rain cape and holding a paper delivery bag stamped "Kitsune 47," mood wistful but resilient, bold flat shapes with crisp linework and subtle grain, reflections in turquoise and vermilion, cinematic backlighting from vending machines, carefully composed in 16:9.
OpenAI: GPT-6 Luna: Model B creates a more evocative, cinematic Osaka side street, with a clear ramen sign, convincing rain-slick reflections, and the correct bag text. Model A better matches the clean vector style and bold poster framing, but its oversized “Kitsune 47” shop sign is less faithful to the requested ramen sign. (Second judge pass, order swapped — scores are the average of both: Model B captures the ramen sign, correctly stamped delivery bag, vending machines, wet Osaka side street, and cinematic turqu | Anthropic: Claude Fable 5.1: Model B nails the key prompt beats: a correctly…
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 conveys dusk with the visible sun and uses the four-color palette consistently, while Model A’s fire includes a bright yellow accent outside the specified palette. Model B also balances the campsite with the mountain range, sun, and trees, creating a more polished overall composition. (Second judge pass, order swapped — scores are the average of both: Model B adheres more closely to the four-color limit and depicts a clear dusk campsite, though the campsite feels small and left-he | Anthropic: Claude Fable 5.1: Model A breaks the four-color restriction with…
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 clearly follows the requested isometric layout, placing the bed along the left wall, the desk beneath the back-wall window, the rug centrally, and the lamp at the front-right. Model B has a polished, inviting look and includes the requested furnishings, but its perspective is not isometric and its composition reads more like a frontal room view. (Second judge pass, order swapped — scores are the average of both: Model B has a warm, polished arrangement that clearly places the bed, center | Anthropic: Claude Fable 5.1: Model A delivers the requested isometric view b…
Midnight Repair Block Party
A coherent multi-subject urban street scene showing five people collaboratively repairing a tiny electric food cart outside a closed laundromat at midnight: one mechanic kneeling by the axle, one woman holding a phone flashlight, a child handing over a crescent wrench, an elderly man balancing a crate of tangerines, and the vendor testing the string lights above the cart; balanced composition with all figures clearly interacting, puddles reflecting magenta storefront glow, illustrated in a polished gouache-and-ink style with layered city signage, 16:9.
OpenAI: GPT-6 Luna: Model A more strongly captures the closed laundromat, vivid magenta reflections, and cohesive gouache-and-ink street scene, though it includes an extra bystander and much of the signage is garbled. Model B clearly depicts the requested interactions, but its laundromat sign says it is open 24 hours and the vendor is awkwardly perched on the cart while fixing the lights. (Second judge pass, order swapped — scores are the average of both: Model A more clearly shows the coordinated repair, phone fla | Anthropic: Claude Fable 5.1: Model B nails nearly every prompt beat—phone fl…
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 A evokes ukiyo-e strongly, but its multiple partially cropped boats and crowded wave forms make the scene less clear. Model B has a more coherent, dramatic composition with a clearly readable fishing boat, stylized foam, and a restrained indigo-and-cream palette. (Second judge pass, order swapped — scores are the average of both: Model B more clearly centers a fishing boat riding a single cresting wave and closely matches the requested muted indigo-and-cream ukiyo-e palette. Model A is str | Anthropic: Claude Fable 5.1: Model A is a near-direct pastiche of Hokusai's…
Matchup powered by OpenRouter.