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
Meta Muse Image Edit and Google’s Nano Banana Pro take on eight image tasks spanning precise object placement, readable text, visual style, and believable hands. The comparison looks at where each model follows the brief closely—and where a polished image still misses a specific requirement.
The saved results are **66.1 for Meta Muse Image Edit and 61.2 for Nano Banana Pro**, with 3 task wins for Meta, 2 for Google, and 3 ties. That edge in aggregate does not settle the matchup: the editorial strength is **Too close to call (not statistical)**, and the saved result remains a tie. Meta’s clearest wins are the watchmaker close-up, Bauhaus juicer ad, and multi-line text poster. It gives the repair scene a more convincing, cluttered bench; makes the juicer read clearly as the advertised appliance; and keeps the requested three-line text layout intact. Those are concrete strengths in composition and brief-following, not just finish. Google wins on exact counting and the melancholy kettle vector. Its seven espresso cups are more clearly separated and evenly arranged, while its kettle scene brings the requested elements together more cohesively. But it splits “NIGHT MARKET” across two lines in the text task, and the kettle image’s brighter, tighter arrangement loses some of the requested contemplative mood—so its wins are specific rather than sweeping. The three ties show why the score gap does not make a convincing verdict: the models trade strengths on the mint-tin still life and library perspective, while neither hands task produces an unambiguous edge. **Final call: a tie—too close to call on this saved task sample.**
Mint Tin Shelf Geometry
A meticulously arranged product still life on a narrow walnut wall shelf, photographed straight-on in soft morning window light, 16:9: a teal mint tin sits exactly between a smoked-glass perfume bottle on the left and a folded mustard-yellow pocket umbrella on the right; a silver fountain pen lies horizontally directly in front of the mint tin; a matte white wireless earbud case is placed behind the umbrella; a tiny potted cactus is on top of a stack of two coral notebooks at the far left; and a brass keyring with three keys hangs below the shelf directly under the perfume bottle.


OpenAI: GPT-6 Luna: Model A more closely matches the requested soft morning still life and correctly places the earbud case behind the umbrella, while Model B places it off to the umbrella’s right. Model B does hang the keys under the perfume bottle, unlike Model A, but Model A’s overall arrangement and lighting are more faithful and polished. (Second judge pass, order swapped — scores are the average of both: Model B more closely matches the requested geometry, especially placing the three-key ring directly beneat | Anthropic: Claude Fable 5.1: Model A nails nearly every spatial relation (ti…
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: Both images show exactly seven individually countable espresso cups in a straight row on white marble. Model B has a cleaner, more evenly spaced presentation, while Model A’s adjacent handles visually run together and its cups sit close to the frame edges. (Second judge pass, order swapped — scores are the average of both: Model B presents exactly seven identical cups in a clean, evenly spaced row, with each cup clearly separate on softly lit white marble. Model A also has seven cups, but the ha | Anthropic: Claude Fable 5.1: Both Model A and Model B render exactly seven e…
Watchmaker Hands Close-Up
A sharply focused half-body portrait of a young watchmaker seated at a cluttered repair bench under a single cool task lamp, realistic editorial photography, 16:9: both hands are fully visible and anatomically correct, with the left hand delicately holding a small brass wristwatch case between thumb and index finger while the right hand uses a fine screwdriver above an open movement tray; all ten fingers are clearly formed and proportionate, wrists natural, shoulders aligned, face turned slightly downward in concentration, with loupe, spring bars, and microfiber cloth scattered around the workstation.


OpenAI: GPT-6 Luna: Model A more fully realizes the cluttered watch-repair bench and clearly shows the case-holding hand and screwdriver working over an open movement tray, with convincing editorial lighting. Model B includes the key actions and tools but feels less cluttered and has a tighter crop that weakens the half-body portrait composition. (Second judge pass, order swapped — scores are the average of both: Model A has a more polished editorial composition and a convincingly detailed, cluttered repair bench, | Anthropic: Claude Fable 5.1: Model A nails the cool single task lamp, a genui…
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 delivers a crisp, photorealistic, symmetrical aisle with clear one-point perspective, warm overhead lighting, and books that consistently diminish with distance. Model B also has strong perspective and an inviting warm atmosphere, but its softer, more stylized rendering is less faithful to the photorealistic request. (Second judge pass, order swapped — scores are the average of both: Model B more clearly delivers the long, empty, warm-lit aisle with strong one-point perspective and consi | Anthropic: Claude Fable 5.1: Model A delivers a clean, photorealistic one-po…
Melancholy Kettle Vector
A clean stylized vector illustration of an electric kettle on a rainy apartment windowsill at blue hour, minimalist but expressive, with smooth shape language, limited palette of slate blue, pale peach, and warm amber, and a quiet melancholic mood; outside the window are soft blurred city lights and diagonal rain streaks, while inside a faint curl of steam rises from the kettle beside a ceramic mug and a tiny packet of instant noodles, composed with generous negative space and crisp edges, 16:9.


OpenAI: GPT-6 Luna: Model A better captures the quiet melancholy through its restrained palette, faint curling steam, and generous negative space, while clearly showing the instant-noodle packet. Model B is polished and readable, but its tighter, brighter arrangement feels less contemplative and the packet is less clearly identifiable. (Second judge pass, order swapped — scores are the average of both: Model B more completely and cohesively delivers the prompt, with a clear kettle, mug, noodle packet, steam, rain, | Anthropic: Claude Fable 5.1: Model B places all requested elements on an actu…
Bauhaus Citrus Juicer Ad
A faithful Bauhaus poster-style illustration of a countertop citrus juicer as the hero object, using strict geometric forms, asymmetric balance, bold primary colors plus black and cream, crisp sans-serif typography blocks, and flat print texture; the juicer is rendered as circles, rectangles, and arcs on a clean cream background with a small measuring cup and two orange halves, lit with the simplified graphic logic of 1920s design rather than realism, composed like a premium appliance advertisement, 16:9.


OpenAI: GPT-6 Luna: Model A clearly presents a countertop citrus juicer with measuring cup and orange halves, using a strong cream, red, blue, black, and yellow Bauhaus poster treatment with legible advertising text. Model B has a fitting geometric palette and vintage feel, but its juicer is much less recognizable and the composition feels sparse and less like a premium appliance advertisement. (Second judge pass, order swapped — scores are the average of both: Model A presents a recognizable geometric citrus juice | Anthropic: Claude Fable 5.1: Model A delivers a fully realized 16:9 Bauhaus…
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.


OpenAI: GPT-6 Luna: Model A follows the requested three-line layout with all text crisp, correctly spelled, and centered on a subtly grained navy background. Model B is also legible and polished, but splits “NIGHT MARKET” across two lines, departing from the requested layout. (Second judge pass, order swapped — scores are the average of both: Model A follows the requested three-line centered layout with crisp, correctly rendered text and a restrained navy poster aesthetic. Model B is legible and polished, but split | Anthropic: Claude Fable 5.1: Model A renders all three lines correctly spell…
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 A more clearly centers the bracelet-making action, with both hands and a plausible thread grip sharply rendered in soft daylight. Model B has an appealing wider setting, but its hand shapes and grip look less natural, and the image is grainier. (Second judge pass, order swapped — scores are the average of both: Model A shows both hands clearly pinching and guiding the threads with a more plausible, readable grip, and its soft daylight and shallow depth of field closely match the prompt. Mo | Anthropic: Claude Fable 5.1: Model B reads as a genuinely candid film-like p…
Matchup powered by OpenRouter.