Head to head: Boogu Image vs Qwen Image 3 Text to Image
Boogu Image vs Qwen Image 3 Text to Image
One model brought occasional flashes of discipline; the other delivered far more complete prompt execution. This matchup turns on consistency across text, counting, layout, and scene-specific details rather than on isolated aesthetic wins.
Qwen Image 3 Text to Image takes this one decisively. The aggregate gap is substantial — 70.3 to 59.6 — and the statistical read is even clearer: a **96% confidence** win, backed by a **6–1 task record with one tie**. That is not a vibes-based edge; it is a broad, repeatable advantage across the benchmark. Where Qwen separated itself was prompt compliance under pressure. It was better at packing in the exact requested details without losing the image: the rust-red tram stop on Vesper Lane with the florist and readable **"Miro 47"** sign; the tiny midnight noodle stand under the elevated rail line with the right wide-angle intimacy and condiment jars; the violinist beside the subway entrance on **Brume Street** with clearer hands, full-body framing, and a believable busking setup. On layout-heavy and counting-heavy prompts, it was simply more reliable: the isometric room placed the bed, desk, rug, and lamp where they were supposed to be, and the lantern-helmet scene actually delivered **13 distinct helmets** instead of falling apart on the core requirement. Boogu Image’s lone win came on the restricted-palette vector task, and it earned it. It stayed truer to the four-color flat-graphic constraint, while Qwen overreached with extra shading and richer detail that weakened adherence. That result says something real about Boogu: when the brief demands restraint and formal discipline over realism or density, it can be the sharper tool. But that strength was too narrow to change the verdict. Even the text poster task ended as a tie, which matters because it denied Boogu an obvious route back into the match. And in several losses, Boogu was not merely less polished — it missed essential prompt anchors: incomplete or inaccurate signage, weaker spatial placement, missing named locations, and in the worst case a complete failure on the counting constraint. Qwen was consistently better at the unglamorous but crucial work of getting the assignment right. **Final call: Qwen Image 3 Text to Image is the stronger image model here, and it wins by being more dependable on the things that actually decide benchmark results — detail fidelity, spatial obedience, text accuracy, and countable object control. Boogu Image has a credible niche in tightly constrained flat-vector work, but Qwen is the better all-around generator.**
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.
Model B better matches the prompt’s minimalist poster feel with cleaner centered hierarchy, a more refined sans-serif treatment, and a subtle grain on a deep navy background while keeping all text crisp and correctly spelled. Model A is also strong and fully legible, but its typography feels heavier and less polished, with spacing and balance that are slightly less elegant than Model B. (Second judge pass, order swapped — scores are the average of both: Model A better matches the prompt with crisp, correctly rendered multi-line text, stronger centered hierarchy, and a cleaner sans-serif treatment. Model B is also good and legible, but its middle and bottom lines are lighter and less crisp, making the overall poster feel slightly less polished and prompt-faithful.)
Golden hour tram stop
A cinematic urban street scene at golden hour: a rust-red tram stop on Vesper Lane with long amber shadows stretching across wet cobblestones, a florist locking up a tiny kiosk under a glowing sign that reads 'Miro 47', pigeons lifting off near the curb, and a lone cyclist passing through the frame; realistic editorial photography style, warm sun skimming building edges, reflective shop windows, moody yet peaceful atmosphere, composed in 16:9.
Model B matches the prompt much more completely, clearly showing a rust-red tram stop on Vesper Lane, a florist locking up beside a readable 'Miro 47' sign, pigeons lifting off, and a lone cyclist in a cinematic golden-hour composition. Model A has strong mood and lighting, but it misses the tram stop emphasis and Vesper Lane text, and the kiosk/sign dominate in a way that weakens prompt adherence. (Second judge pass, order swapped — scores are the average of both: Model B matches more of the prompt simultaneously: a rust-red tram stop on Vesper Lane, a florist locking up beside a clearly rendered 'Miro 47' sign, pigeons lifting off, a lone cyclist, wet cobblestones, and strong golden-hour reflections in a cinematic editorial composition. Model A has appealing light and mood, but it misses the tram-stop emphasis and florist-locking-up action, and the text rendering is less accurate with 'Miiro 47' and no visible Vesper Lane signage.)
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.
Model A adheres more strictly to the restricted four-color flat-vector requirement, with a clean balanced campsite composition and no obvious off-palette intrusions. Model B is visually richer, but it introduces extra shades and more complex lighting/details that weaken prompt adherence to the strict palette and flat simplicity. (Second judge pass, order swapped — scores are the average of both: Model A adheres more cleanly to the restricted four-color flat-vector requirement, with a simpler, more cohesive palette and no obvious off-palette accents. Model B is more detailed and visually engaging, but it appears to introduce extra shades and minor off-palette variations, which weakens prompt adherence despite solid composition.)
Wide-angle noodle stand
A wide-angle street photography scene shot from waist height with a 24mm lens effect, showing a tiny midnight noodle stand tucked under an elevated rail line on Calder Row, exaggerated perspective lines pulling the eye toward the vendor, nearby stools and condiment jars appearing larger in the foreground while neon signs recede dramatically into the distance; documentary-realistic style, glossy rain-slick asphalt, cool cyan and magenta night lighting, dynamic framing in 16:9.
Model B adheres more closely to the prompt by clearly presenting a tiny midnight noodle stand under an elevated rail line on Calder Row, with strong wide-angle foreground exaggeration, visible condiment jars, and convincing documentary-realistic detail. Model A has appealing cyan-magenta mood and solid perspective, but it feels more distant and generic, with weaker foreground emphasis and less specific prompt detail than Model B. (Second judge pass, order swapped — scores are the average of both: Model B better matches the prompt’s tiny midnight noodle stand under an elevated rail line, with stronger foreground emphasis on stools and condiment jars, convincing wide-angle perspective, and clearer documentary realism. Model A has excellent cyan-magenta mood and strong leading lines, but it feels more staged and misses key prompt details like prominent condiment jars and the tucked-in, waist-height street-photography intimacy.)
Street violinist with hands
A full-body portrait of a young street violinist busking beside a subway entrance on Brume Street, shown clearly from head to toe with natural anatomy and both hands fully visible: one hand pressing the violin strings and the other holding the bow correctly, all fingers distinct and accurately formed, relaxed shoulders, believable posture, proportional limbs, open violin case with scattered coins, contemporary realistic digital painting style, soft overcast daylight, detailed urban background without obscuring the figure, 16:9.
Model B adheres more closely to the prompt by explicitly placing the violinist beside a subway entrance on Brume Street, showing the full body clearly, and including an open case with scattered coins in a convincing urban scene. Model A is attractive and technically solid, but it misses the Brume Street cue, the case is less integrated as a busking setup, and the hand/pose depiction is slightly less clear than Model B's. (Second judge pass, order swapped — scores are the average of both: Model B adheres more closely to the prompt with a clearer Brume Street sign, full-body framing, visible hands, and a convincing open case with scattered coins in a realistic painted style. Model A has strong centered composition and pleasing atmosphere, but the street-name text is incomplete, the case/coins are less emphasized, and the anatomy/hand clarity is slightly less convincing.)
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.
Model B is more faithful to the ukiyo-e woodblock print prompt, with flatter color planes, bolder outlines, more stylized foam, and a stronger traditional composition centered on a fishing boat riding a cresting wave. Model A is attractive and evocative, but it feels more simplified and less authentically woodblock-like, with a looser composition and less convincing traditional detailing. (Second judge pass, order swapped — scores are the average of both: Model B better matches the prompt with a clear ukiyo-e woodblock look, bold outlines, stylized foam, a fishing boat actively riding a cresting wave, and a muted indigo-and-cream dawn palette. Model A is attractive and evocative, but it departs more from the prompt through a less faithful boat design and composition that feels more like a general Japanese-inspired seascape than a traditional ukiyo-e fishing-boat scene.)
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.
Model B better matches the requested spatial layout: the bed is clearly against the left wall, the desk sits under the back-wall window, the round rug is centered on the floor, and the floor lamp is placed in the front-right corner. Model A has a cleaner isometric/vector look, but its bed is not clearly against the left wall and the lamp reads more as mid-right than distinctly front-right. (Second judge pass, order swapped — scores are the average of both: Model B follows the requested layout more closely, with the bed on the left wall, the desk under the back-wall window, and the floor lamp clearly placed in the front-right corner; its flat-vector isometric style is also more polished and visually appealing. Model A captures the basic room elements and centered rug, but the lamp sits along the right wall rather than the front-right corner, and the overall composition feels sparser and less refined.)
Thirteen lantern helmets
A minimalist line art illustration of a narrow back-alley street market called 'Copper Finch Arcade', showing EXACTLY 13 distinct vintage scooter helmets hanging from a wire between two fire escapes, each helmet fully visible and individually countable, with a noodle cart below, cracked pavement, utility cables, and a small painted wall ad reading 'Kestrel Ink'; clean monochrome lines with a few muted accent colors, straight-on composition, crisp separation between every object.
Model B adheres much more closely to the prompt by showing exactly 13 fully visible, distinct vintage scooter helmets with clear separation, a straight-on alley composition, cracked pavement, utility cables, and legible wall text including 'Kestrel Ink' and 'Copper Finch Arcade'. Model A is cleaner but fails the core counting requirement with only 6 helmets and omits the 'Copper Finch Arcade' market name, making it substantially less compliant despite decent line quality. (Second judge pass, order swapped — scores are the average of both: Model B adheres much more closely to the prompt by showing exactly 13 fully visible, individually countable helmets, the noodle cart, cracked pavement, utility cables, fire escapes, and a readable 'Kestrel Ink' wall ad in a clean straight-on composition; its main miss is replacing the alley name with a large hanging sign rather than integrating it as the market name in-scene. Model A has a pleasing minimalist layout but fails the core counting requirement with far fewer than 13 helmets and omits the 'Copper Finch Arcade' text entirely, making it substantially less prompt-faithful.)
Matchup powered by OpenRouter.