Head to head: AuraFlow vs Fibo Bbq Preview
AuraFlow vs Fibo Bbq Preview
By Ryan Merket · Published
A comparison of two image models across exact-text rendering, prompt constraints, composition, style, and detailed scene-building. The results turn less on polish alone than on which model reliably follows the brief.
Fibo Bbq Preview takes this matchup decisively: it scored 60.0 to AuraFlow’s 47.3, with a tentative lean verdict. Across the task results, Fibo won 12, AuraFlow won 3, and one was a tie. This was not a narrow edge built on one standout image; Fibo repeatedly got more of the requested scene right. Its clearest advantage was instruction-following under constraints. Fibo avoided forbidden objects in the reading-nook and no-umbrella prompts, and it placed the bedroom furniture more faithfully. It also delivered the requested croquembouche and amber sugar ribbon, and more consistently matched the ukiyo-e prompt’s muted indigo-and-cream palette and dawn setting. AuraFlow’s empty bookstore aisle was a genuine win: its clean, centered perspective beat Fibo’s stool-cluttered version. AuraFlow could be the more polished image-maker in individual cases. Its retro soda poster had stronger illustration and convincing bottle-and-orange imagery, and it also performed well on the rooftop brunch scene. But those strengths came with recurring misses against the brief: garbled or misspelled lettering, unwanted umbrellas or plants, and misplaced room elements. Fibo had its own flaws—notably stray lettering on the soda poster and occasional scene details that drifted—but its errors more often left the core request intact. **Verdict: Fibo Bbq Preview wins clearly.** AuraFlow has flashes of better craft, but Fibo is the more dependable model when the prompt’s specific details and hard constraints matter.
Mirelle Soda Poster
A crisp graphic poster photographed straight-on in a boutique grocery window, designed in elegant retro typography and easy-to-read lettering: a cream background with cobalt and coral accents, an illustrated glass bottle with condensation and sliced blood orange, and the clearly legible text exactly reading "MIRELLE SODA" on the top line and "SPARKLING QUINCE" below it; clean mid-century design, sharp focus, even daylight, uncluttered composition.
OpenAI: GPT-6 Luna: Model B clearly renders both required lines exactly and uses the requested cream, cobalt, and coral palette, though its extra gibberish lettering makes the poster cluttered. Model A has a more polished retro illustration and convincing bottle-and-orange imagery, but misspells both required text elements. (Second judge pass, order swapped — scores are the average of both: Model B more closely matches the required wording and color palette, though its subtitle appears misspelled and the extra lett | Anthropic: Claude Fable 5.1: Model B renders the key headline 'MIRELLE SODA'…
Rooftop Brunch Orchestra
A whimsical claymation-style 16:9 rooftop brunch scene at sunrise, composed as a balanced ensemble image: a baker setting down a tower of cardamom waffles, a trumpet player in a teal suit serenading two friends pouring guava mimosas, a small terrier on a stool watching a spinning platter of figs, and a florist weaving marigolds through the backs of four mismatched chairs; everyone is interacting naturally around one long table with copper coffee pots, apricot jam jars, and striped plates, lit by soft golden morning light and framed from a slightly elevated angle.
OpenAI: GPT-6 Luna: Model B better captures the rooftop sunrise, marigolds, figs, waffle tower, and whimsical claymation feel, with a lively, cohesive composition. Model A includes several requested elements too, but its rooftop setting is less clear and its heavy border and sparse interaction make it feel less complete. (Second judge pass, order swapped — scores are the average of both: Model B more vividly captures the whimsical rooftop sunrise, claymation look, towering waffles, copper pots, marigolds, and dog w | Anthropic: Claude Fable 5.1: Model B captures more of the prompt's specifics…
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 includes a plant, a lamp, and framed artwork, directly violating all three explicit exclusions. Model B provides the armchair, stacked books, steaming mug, window, and warm light without any of the forbidden items, and is also polished and inviting. (Second judge pass, order swapped — scores are the average of both: Model B closely follows the requested nook and warm afternoon light while avoiding plants, lamps, and artwork; its main shortcoming is the square format rather than 16:9. Mod | Anthropic: Claude Fable 5.1: Model A violates every negation in the prompt —…
Perspective & scale
A photorealistic one-point-perspective shot down a long empty library aisle, shelves receding to a single vanishing point, books getting consistently smaller with distance and nothing warped or wrongly sized, warm overhead light, 16:9.
OpenAI: GPT-6 Luna: Model A closely matches the empty, photorealistic one-point-perspective aisle, with shelves and books receding consistently toward a centered vanishing point. Model B has striking warm lighting and strong perspective, but the prominent stool breaks the requirement for an empty aisle. (Second judge pass, order swapped — scores are the average of both: Model A better fulfills the empty-aisle requirement and presents a clean, convincing one-point perspective with consistent recession. Model B has r | Anthropic: Claude Fable 5.1: Model A delivers a cleaner, more consistent one…
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 desk directly beneath the back-wall window and keeps the lamp in the front-right corner, while Model A’s desk is not beneath a window. Model B also has a cleaner, more cohesive flat-vector style, though its rug is partly obscured by the bed. (Second judge pass, order swapped — scores are the average of both: Model B places all four requested elements in the specified locations and maintains a clean, consistent isometric flat-vector style. Model A has a desk away from the windo | Anthropic: Claude Fable 5.1: Model B places every element as specified—bed o…
Rainy Dumpling Alley No Umbrellas
A cinematic 16:9 nighttime food-stall alley in a drizzly old port district, rendered as detailed live-action photography with glossy reflections on wet cobblestones and warm lantern light; steam rises from a jade-green dumpling cart run by a vendor in a saffron apron while customers huddle under awnings, neon signs shimmer in puddles, crates of purple cabbages and brass teapots line the lane, and the entire scene must contain absolutely NO umbrellas anywhere.
OpenAI: GPT-6 Luna: Model B captures the steaming green dumpling cart, saffron apron, customers, wet reflective alley, purple cabbages, and brass teapots without visible umbrellas. Model A has strong atmosphere and detail, but prominently includes umbrella-like canopies over the cart and customers, directly violating the prompt. (Second judge pass, order swapped — scores are the average of both: Model B closely matches the requested scene, including the steaming jade-green cart, saffron apron, wet reflective lane, | Anthropic: Claude Fable 5.1: Model A violates the explicit hard constraint by…
Sugar Ribbon Pour
An ultra-detailed stop-motion-inspired 16:9 dessert action shot: a pastry chef in a flour-dusted charcoal jacket mid-swing, pulling a glowing ribbon of molten amber sugar through the air over a black slate counter while a half-finished croquembouche tilts beside bowls of pistachios and tiny silver tongs; capture the sugar arc frozen at peak motion with flying sparks, scattered crumbs, dramatic side lighting, shallow depth of field, and strong diagonal composition that conveys speed and precision.
OpenAI: GPT-6 Luna: Model B better matches the prompt with a recognizable croquembouche, flour-dusted charcoal jacket, pistachios, and a glossy amber sugar ribbon. Model A has a more dramatic sweeping arc, but misses key details, especially the croquembouche, and its sugar reads more like sparkling powder. (Second judge pass, order swapped — scores are the average of both: Model B closely matches the molten amber sugar ribbon, flour-dusted charcoal jacket, croquembouche, pistachios, slate counter, and dramatic ligh | Anthropic: Claude Fable 5.1: Model B nails the key prompt elements—flour-dus…
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 captures the dramatic ukiyo-e wave and boats, but its bright turquoise and orange palette is less faithful to the requested muted indigo and cream, and dawn is not clearly conveyed. Model B combines a cresting wave, fishing boat, sunrise, and restrained indigo-and-cream woodblock styling in a cohesive composition. (Second judge pass, order swapped — scores are the average of both: Model B better matches the muted indigo-and-cream palette and clearly conveys dawn with the rising sun, whil | Anthropic: Claude Fable 5.1: Model A closely mimics Hokusai's Great Wave wit…
Matchup powered by OpenRouter.