How to Prompt Qwen Image for Photorealism (2026): Formulas + Examples

Qwen Image 3.0 reads long, brief-style prompts, so photorealism comes from concrete camera and lighting language, not the phrase "make it realistic". Describe a specific photo — a subject, a lens, a named light source, a setting — and the model builds a plausible frame around it. This guide gives you a 6-part formula, the settings that matter, and 10 copy-paste prompts you can adapt; for more starting points see the best Qwen Image prompts roundup and the Qwen Image prompt cheat sheet.

The 6-part photorealism formula

Every photorealistic Qwen prompt answers six questions in order: who or what, shot with what, lit how, where, in what medium, and within what limits. Fill all six and the model has nothing left to guess.

1. Subject — the exact thing in frame and what it is doing: "a 60-year-old fisherman mending a net", not "a person".
2. Camera and lens — a real body and focal length plus aperture: "shot on a 50mm lens at f/2.0", which sets perspective and depth of field.
3. Lighting — a named setup and direction: "soft overcast daylight from camera left", "golden hour backlight", "single softbox key".
4. Setting — the environment and a few grounding details: "a weathered wooden dock at dawn, coiled ropes, wet planks".
5. Style and medium — say it is a photograph and name a genre: "documentary photograph", "editorial portrait", "product photo".
6. Constraints — a final line with aspect ratio and resolution so the canvas is fixed: "Aspect ratio 3:2, 2K."

Here is the copyable skeleton. Swap the bracketed parts:

A [documentary/editorial/product] photograph of [subject doing specific action], [key physical details]. Shot on a [focal length] lens at [f-stop], [shallow/deep] depth of field. Lit by [named light source] from [direction], [quality: soft/hard, warm/cool]. Setting: [environment with 2-3 grounding details]. Natural skin and material texture, true-to-life color, realistic contrast. Aspect ratio [3:2], 2K.

And the same skeleton filled in for one shot:

A documentary photograph of a 60-year-old fisherman mending a green net with weathered hands, salt-stained yellow raincoat, gray stubble. Shot on a 50mm lens at f/2.0, shallow depth of field with a softly blurred harbor behind him. Lit by soft overcast daylight from camera left, cool and even. Setting: a weathered wooden dock at dawn, coiled ropes, wet planks, a moored trawler out of focus. Natural skin texture with visible pores, true-to-life color, realistic contrast. Aspect ratio 3:2, 2K.

Why it works: the six slots give Qwen a subject, optics, light direction, place, medium, and canvas — leaving no room for the guesswork that produces flat, illustrated results.

Camera and lens language

Lens and f-stop language tells Qwen how the frame should look optically: focal length sets perspective and compression, aperture sets depth of field and background blur. Name both.

A wide lens (24-35mm) exaggerates space and keeps most of the scene sharp — good for interiors and landscapes. A standard lens (50mm) looks natural and human. A short telephoto (85-135mm) compresses features flatteringly and is the classic portrait choice. Aperture is the depth-of-field dial: f/1.4-f/2.8 gives a shallow plane and creamy background blur (bokeh); f/8-f/16 keeps foreground and distance both sharp. Add words like "shallow depth of field", "background falls out of focus", or "everything in focus front to back" so the intent is unmistakable.

An editorial portrait of a woman with freckles and short auburn hair, looking just off camera. Shot on an 85mm lens at f/1.8, very shallow depth of field, only the eyes tack-sharp and the background melting into soft bokeh. Lit by a large softbox key from camera right with a subtle fill. Setting: a neutral gray studio backdrop. Natural skin texture with fine pores, realistic catchlights in the eyes. Aspect ratio 4:5, 2K.

Best for: flattering headshots where you want the subject isolated from the background.

Lighting that reads as real

Real photos are defined by where the light comes from and how soft it is, so name an actual setup rather than saying "well lit". Direction plus quality does most of the work.

Useful named setups: golden hour (warm, low, long shadows), overcast daylight (soft, shadowless, forgiving), window light (soft directional from one side), softbox key with fill (studio portrait), hard midday sun (crisp shadows, high contrast), Rembrandt lighting (a small triangle of light on the shadow-side cheek), and rim/backlight (a bright edge separating subject from background). Always state the direction — "from camera left", "backlit" — and the quality — soft or hard, warm or cool.

A documentary photograph of a barista pulling an espresso shot, steam rising from the group head, focused expression. Shot on a 35mm lens at f/2.8. Lit by warm window light from camera left with a soft rim of afternoon sun catching the steam. Setting: a small specialty coffee bar, brass machine, wooden counter, blurred shelves behind. Natural skin tones, realistic reflections on the metal, true color. Aspect ratio 3:2, 2K.

Why it works: naming a direction and a warm quality gives the scene believable shadows and highlights instead of flat, even AI lighting.

Settings: aspect ratio, resolution, guidance

Qwen has no --flags; you set the frame with a UI preset or plain language, and a couple of advanced dials fine-tune realism. Match the ratio to the shot and keep guidance moderate.

Aspect ratios (supported: 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3): use 3:2 or 16:9 for landscapes, cars, and cinematic scenes; 2:3 or 3:4 for portraits and fashion; 1:1 for product and food hero shots; 9:16 for phone-native vertical. Resolution: output at 2K (up to 2048x2048) for the sharpest texture; a 3:2 frame lands around 1536x1024. Guidance (true_cfg_scale): keep it near the default of ~4 for photorealism — lower (2-3) is softer and more natural but drifts from the prompt, higher (6-10) over-enforces and can look glossy and CGI-like. Inference steps: more steps add fine detail at the cost of speed; the default is fine, bump it up only for high-texture close-ups.

10 photorealistic example prompts

Ten complete, copy-paste prompts across the genres that come up most. Each ends with a constraints line — swap subjects freely and keep the camera and lighting structure intact.

1. Editorial portrait

An editorial portrait of a man in his 40s with a salt-and-pepper beard and a charcoal wool coat, calm confident gaze slightly off camera. Shot on a 105mm lens at f/2.0, shallow depth of field, background softly blurred. Rembrandt lighting from a softbox at camera right, gentle fill on the shadow side. Setting: a moody dark studio with faint texture on the backdrop. Natural skin texture with visible pores and fine lines, realistic catchlights. Aspect ratio 4:5, 2K.

Best for: magazine-style headshots and author photos with dramatic, flattering light.

2. Street candid

A candid street photograph of a young woman laughing while crossing a rain-slicked crosswalk, motion in her step, umbrella half-open. Shot on a 35mm lens at f/2.8, mild motion in the background. Lit by overcast afternoon daylight with neon shopfront reflections on the wet asphalt. Setting: a busy city intersection, blurred pedestrians, yellow taxi out of focus. Natural skin tones, realistic wet-surface reflections, true-to-life color. Aspect ratio 3:2, 2K.

Best for: documentary, lifestyle, and editorial-city imagery with authentic energy.

3. Product hero

A product hero photograph of a matte-black stainless steel water bottle standing upright, condensation beads on the surface. Shot on a 100mm macro lens at f/8, sharp front to back. Lit by a large softbox from upper left with a subtle gradient reflection and a second fill card on the right. Setting: a seamless light-gray studio sweep, soft contact shadow beneath the bottle. Crisp material texture, accurate metal reflections, clean color. Aspect ratio 1:1, 2K.

Best for: e-commerce and ad hero shots — see more in product photography prompts.

4. Food

An overhead food photograph of a bowl of ramen with a soft-boiled egg, nori, scallions, and glistening broth, steam rising. Shot on a 50mm lens at f/4, slight depth. Lit by soft window light from the top of frame with a fill card below to lift shadows. Setting: a dark textured stone table, chopsticks resting on a linen napkin, a small side dish. Appetizing natural color, realistic broth sheen, crisp garnish detail. Aspect ratio 1:1, 2K.

Best for: menu, recipe, and delivery-app imagery that looks freshly shot.

5. Interior

An architectural interior photograph of a sunlit Scandinavian living room, oak floors, a linen sofa, a large fiddle-leaf plant. Shot on a 24mm lens at f/8, everything sharp front to back, straight verticals. Lit by soft daylight pouring through floor-to-ceiling windows on the left, gentle bounce fill. Setting: minimalist decor, a ceramic mug on the coffee table, morning light and long soft shadows. True-to-life color, realistic wood grain and fabric texture. Aspect ratio 3:2, 2K.

Best for: real-estate, interior-design, and hospitality visuals.

6. Landscape

A landscape photograph of a misty pine valley at golden hour, layered ridgelines fading into haze, a river catching warm light. Shot on a 35mm lens at f/11, deep depth of field, sharp foreground rocks to distant peaks. Lit by low warm sun from camera right, long soft shadows, atmospheric depth. Setting: wet foreground grasses, a lone hiker for scale, drifting mist between the trees. Natural color, realistic aerial haze, fine detail. Aspect ratio 16:9, 2K.

Best for: travel, editorial, and print-worthy scenic work.

7. Automotive

An automotive photograph of a deep-blue sports coupe on a coastal mountain road at dusk, three-quarter front angle. Shot on a 70mm lens at f/5.6, sharp car with gently blurred road behind. Lit by fading golden-hour light from camera left with a cool sky gradient reflected on the paint and glass. Setting: guardrail, ocean far below, motion-blurred asphalt. Realistic paint reflections and metallic flake, true color, believable contrast. Aspect ratio 16:9, 2K.

Best for: car ads, dealership visuals, and hero banners.

8. Wildlife / macro

A macro wildlife photograph of a dew-covered dragonfly perched on a reed at dawn, wings backlit and translucent. Shot on a 100mm macro lens at f/4, razor-thin focus on the eyes and body, background dissolving into soft green bokeh. Lit by soft backlight from behind catching every water droplet. Setting: a quiet pond edge, blurred reeds, cool morning tones. Extreme fine detail, realistic droplet refraction, natural color. Aspect ratio 3:2, 2K.

Best for: nature close-ups and hero macro shots with dramatic focus.

9. Fashion

A fashion editorial photograph of a model in a flowing red silk dress caught mid-stride, fabric in motion. Shot on a 85mm lens at f/2.5, sharp on the face and dress, background softly compressed. Lit by hard directional afternoon sun creating crisp shadows, a bounce fill to open the face. Setting: a sun-washed concrete plaza, minimal background, strong graphic light. Natural skin texture, realistic fabric sheen and folds, saturated but true color. Aspect ratio 2:3, 2K.

Best for: lookbooks, campaigns, and editorial fashion spreads.

10. Night / neon

A night street photograph of a rain-soaked alley lit by pink and blue neon signage, reflections pooling on the pavement. Shot on a 35mm lens at f/1.8, shallow depth of field, glowing bokeh from distant signs. Lit only by mixed neon and a single warm streetlamp, moody high-contrast color. Setting: a narrow city lane, wet cobblestones, a lone figure with an umbrella walking away. Realistic wet reflections, natural neon color bleed, film-like grain. Aspect ratio 9:16, 2K.

Best for: cinematic, cyberpunk-leaning night scenes and vertical social posts.

Mistakes to avoid

Most flat or fake-looking results trace back to a handful of habits. Fix these first.

• Tag soup. Comma-separated keyword piles ("woman, beautiful, 8k, ultra detailed, masterpiece") confuse Qwen — it wants a described scene, not a tag list.
• Vague "realistic". The word alone does little. Concrete camera, lens, and lighting language is what actually produces realism.
• Conflicting styles. Words like cinematic, concept art, artstation, or vibrant pull toward rendering — one of them can flip a photo into an illustration.
• Over-long headlines in a photo. If your shot contains a sign or label, keep the quoted text tight; a wall of words in a photorealistic frame breaks the illusion.
• Ignoring lighting. No named light source means flat, directionless AI lighting. Always state where the light comes from and how soft it is.
• Wrong aspect ratio. A portrait squeezed into 16:9 or a landscape crammed into 1:1 reads as off. Match the ratio to the subject in the constraints line.