Head to head: Meta Muse Image Edit vs Google: Nano Banana Pro (Gemini 3 Pro Image)
Meta Muse Image Edit vs Google: Nano Banana Pro (Gemini 3 Pro Image)
By Ryan Merket · Published
A comparison of two image models across exact-count food styling, reflective surfaces, palette constraints, spatial layouts and richly specified scenes.
Meta Muse Image Edit takes this matchup on both the aggregate score—66.9 to Nano Banana Pro’s 59.1—and the task count, winning six of eight prompts. The limited confidence rating makes this a clear result, not a squeaker, though Google lands two meaningful wins of its own. Muse’s edge is strongest when the prompt depends on specific visual details. It gets the petit-fours count right, comes closer on the demanding reflections brief, and more faithfully handles the tea spread’s specified vessels and ingredients. It also wins the extreme yuzu-glaze macro, where Google’s wider framing misses the point of the shot, and produces the stronger night-market scene. Google’s wins are instructive: its restricted-palette campsite feels more complete and cohesive, and its bedroom layout makes the requested furniture placement clearer. Those are real strengths in composition and scene organization, but they don’t outweigh Muse’s consistency across the food, still-life and detail-heavy prompts. In the reading-nook task, both models respect the key exclusions; Muse’s warmer, cleaner presentation gives it the nod. **Final call: Meta Muse Image Edit wins, 6–2.** It is the more dependable choice when a prompt stacks exact objects, counts, materials or close-up details—and Google’s two wins show it remains competitive when composition and spatial clarity take priority.
Thirteen petit fours
Top-down flat lay of exactly thirteen distinct petit fours on pale pistachio parchment, each piece separated and fully visible for easy counting: 2 raspberry opera bars, 3 lemon-glaze cubes, 1 blackberry tartlet, 2 pistachio domes, 1 coffee éclair, 2 apricot diamonds, and 2 chocolate mirror squares; minimalist patisserie styling, even diffused lighting, precise graphic composition, 16:9.
OpenAI: GPT-6 Luna: Model A shows the requested assortment with exactly thirteen pieces, including three lemon-glaze cubes, while Model B appears to include four lemon cubes. Model B has a slightly more orderly layout, but Model A’s clear spacing and polished presentation make it the closer match overall. (Second judge pass, order swapped — scores are the average of both: Model B presents a clean, well-spaced flat lay with every pastry fully visible, but includes four lemon cubes instead of three. Model A gets the | Anthropic: Claude Fable 5.1: Model A delivers exactly thirteen pieces with th…
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 is a more compelling close-up, with dramatic side light and clear window and apple reflections across the chrome and polished counter; Model B is attractive but framed wider and its water surface shows fewer legible reflections. Neither image clearly renders both required reflections on the water’s surface, though Model A comes closer overall. (Second judge pass, order swapped — scores are the average of both: Model A more clearly shows the window and red apple reflected in the chrome te | Anthropic: Claude Fable 5.1: Model A delivers the fullest take on the brief:…
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 creates a more fully realized dusk campsite, with a lake, campers, tents, and layered mountains arranged in a lively, balanced scene while keeping to the four-color flat-vector palette. Model A also follows the palette and style, but its oversized foreground tents and sparse campsite details make the composition feel less cohesive. (Second judge pass, order swapped — scores are the average of both: Model B presents a more complete dusk campsite, with mountains, tents, a lake, trees, and | Anthropic: Claude Fable 5.1: Model A strictly honors the four-color restricti…
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 has a warm, polished reading-nook composition with all requested objects and no lamps or artwork, though foliage is visible outside the window. Model B also includes the requested objects but shows more prominent garden plants through the window and has a darker, grainier finish. (Second judge pass, order swapped — scores are the average of both: Both images include the requested reading nook elements and warm window light, with no visible indoor plants, lamps, or artwork. Model A presen | Anthropic: Claude Fable 5.1: Both Model A and Model B satisfy the negations…
Night market broth stall
Lively illustrated night-market food scene with several interacting subjects in one balanced composition: a noodle vendor ladling saffron broth from a brass pot, a teenager handing over a striped bowl, an elderly woman seasoning dumplings with chili oil, and a child on a stool fanning steam toward hanging paper lanterns, while baskets of herbs, stacked ceramic spoons, and a chalkboard menu frame the group; cinematic warm rain-slick lighting, cohesive storytelling, detailed contemporary gouache style, 16:9.
OpenAI: GPT-6 Luna: Model A includes all four interactions and framing details in a richly layered, rain-lit market scene, with especially strong detail and legible menu text. Model B is charming and follows the prompt well, but its composition is less intricate and some menu text is less clear. (Second judge pass, order swapped — scores are the average of both: Model B more closely matches the lively gouache brief, with all four interactions clearly staged and the child visibly fanning steam beneath the lanterns. | Anthropic: Claude Fable 5.1: Model A delivers every requested element with ri…
Jacquard tea spread
Crisp high-end studio still life of a tea service arranged on an indigo jacquard tablecloth, emphasizing ultra-fine texture and detail: the woven fabric pattern, hairline crazing in a celadon teacup, brushed metal infuser, condensation on a glass pitcher, and serrated edges of fresh shiso leaves all rendered cleanly without smudging; soft directional daylight, restrained color palette, product-photography realism, 16:9.
OpenAI: GPT-6 Luna: Model A more completely matches the requested still life, including a celadon cup and saucer, condensation-covered pitcher, brushed infuser, loose tea, serrated shiso, and richly textured indigo jacquard. Model B has appealing fabric detail, but the prominent cropped teapot and handle-less bowl diverge from the specified tea service and weaken the composition. (Second judge pass, order swapped — scores are the average of both: Model B renders the jacquard weave and shiso edges especially clearly | Anthropic: Claude Fable 5.1: Model A delivers a clean, fully framed studio s…
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 rug more centrally and makes the floor lamp’s front-right position especially clear, while keeping the desk beneath the back-wall window and the bed along the left wall. Model A is also polished and well arranged, but its rug sits farther toward the foreground and its lamp is less clearly in the corner. (Second judge pass, order swapped — scores are the average of both: Model B places the rug more centrally and clearly satisfies the requested arrangement, with crisp, consisten | Anthropic: Claude Fable 5.1: Model B places every element where the prompt s…
Macro yuzu glaze
Extreme macro close-up food photograph of a single yuzu-glazed canelé resting on a black slate chip, shot at table level with a 100mm macro lens so the ridged caramelized crust fills the frame, the front edge tack-sharp and the background melting into creamy bokeh; tiny beads of citrus glaze, flecks of candied peel, and a few drifting sugar crystals visible under moody side lighting, editorial studio realism, 16:9.
OpenAI: GPT-6 Luna: Model A more closely delivers the requested extreme macro, with sharply detailed ridges, citrus glaze, peel, sugar crystals, slate, and creamy bokeh under dramatic side lighting. Model B shows the whole pastry clearly, but its wider framing and less pronounced macro detail miss the prompt’s defining close-up effect. (Second judge pass, order swapped — scores are the average of both: Model A better captures the requested extreme macro, with dramatic side lighting, visible citrus peel and sugar cr | Anthropic: Claude Fable 5.1: Model A delivers the requested extreme macro wi…
Matchup powered by OpenRouter.