👤

GPT Image 2 Portrait Prompts

GPT Image 2 portrait prompts work best when they describe a photography decision, not just a person. State the portrait task, define a fictional adult subject, explain where the image will be used, choose a photographic style, specify camera position and crop, design the light, name visual failures to avoid, and request a concrete output size. The 15 copy-ready examples below cover neutral headshots, environmental founder and creator profiles, window light, golden hour, monochrome documentary work, low-key and colored-gel lighting, realistic skin texture, mature adults, glasses, curls, very deep skin exposure, an authorized background edit, and a three-composition identity test. Each visible prompt is paired with its own reviewed GPT Image 2 result. Replace the fictional subject details and context for your project, but keep the composition, lighting, constraints, and output clauses until you have a clean baseline. Then adjust one variable at a time. Rerun any result that smooths away real texture, changes identity, hides the eyes, clips skin highlights, invents brands or text, or resembles a real public figure.

Copy tested GPT Image 2 portrait prompts for professional headshots, editorial profiles, natural skin, difficult lighting, age fidelity, glasses, curls, deep skin, and identity-preserving edits.

15 prompts · 15 reviewed GPT Image 2 results · Updated 2026-07-14

Choose the right GPT Image 2 portrait workflow

Use a neutral headshot when identity and trust are the only story, an environmental portrait when place explains the person, and a controlled-light portrait when mood or a technical challenge is the teaching point. Keep one primary goal per generation. A clear crop, one lighting plan, and explicit texture or identity constraints produce a result that is easier to inspect and correct.

Four steps for reliable GPT Image 2 portraits

  1. Define the useChoose a staff headshot, editorial profile, location portrait, lighting study, or rights-cleared edit before describing wardrobe and scenery.
  2. Lock subject anchorsDescribe an original adult, apparent age, skin tone, hair, wardrobe, expression, and the few identity details that must remain stable.
  3. Design camera and lightName crop, perspective, camera height, subject placement, key direction, fill, catchlights, white balance, and background separation.
  4. Inspect before reuseCheck eyes, hands, skin texture, hair edges, glasses, highlights, text, brands, identity, and public-figure resemblance; rerun targeted failures.

Portrait prompt parameter guide

GoalLayoutBest forAdjustAvoid
Professional headshotChest-up, eye level, simple backdropBios, teams, speakersLens, expression, backdrop warmthBeauty-filter skin and tight shoulder crops
Environmental portraitWaist-up with controlled contextFounder and creator storiesWorkplace, props, camera distanceBusy settings that compete with the face
Lighting studyHead-and-shoulders or chest-upWindow, low-key, gels, skin exposureKey size, fill ratio, color, rimHidden eyes, clipped highlights, color casts
Identity-safe editMatch the authorized source cropBackground replacementNew setting and depth of field onlyRedrawing the face, pose, wardrobe, or age

Failure corrections before you rerun

Skin looks plastic

Why it happens: The prompt asks for polish but does not protect natural texture.

Prompt correction: Add ‘preserve pores, peach fuzz, fine lines, under-eye texture, facial asymmetry, and realistic highlight rolloff; no beauty filter.’

The eyes are hidden

Why it happens: Low-key light, glasses, or hair blocks the strongest identity anchors.

Prompt correction: Require both eyes to remain readable, place catchlights deliberately, feather softboxes away from glasses, and add minimal neutral fill.

Deep skin turns ashy or orange

Why it happens: Exposure and white balance are described vaguely while colored light or warm grading dominates.

Prompt correction: Request calibrated neutral white balance, broad neutral fill, controlled highlights, and true warm-neutral undertones without gray lift or red saturation.

The edited portrait becomes a different person

Why it happens: The request describes a new scene but does not freeze the authorized subject.

Prompt correction: State that only the background may change and list the face, hair, expression, pose, crop, clothing, skin exposure, and catchlights that must remain identical.

15 copy-ready GPT Image 2 portrait prompts

Every visible prompt below generated the output shown in its card. Copy the full instruction, replace only rights-cleared subject and context details, and keep the lighting, constraints, and output clauses until your first result passes review.

Portrait · GPT Image 2 · Prompt 01

Neutral Professional Headshot

For staff bios, speaker pages, and professional profiles that need a clean, trustworthy portrait.

GPT Image 2 portrait prompts result showing a neutral professional headshot
Neutral Professional Headshot | Best for: For staff bios, speaker pages, and professional profiles that need a clean, trustworthy portrait. | Prompt ENP04-P01 | Check: skin texture, eye sharpness, shoulder room, catchlights, and neutral color | If it fails: If the face looks over-retouched, add ‘retain pores, fine lines, facial asymmetry, and natural under-eye texture; no beauty filter.’
Expected result

A polished chest-up headshot with believable skin texture, clean separation, and a calm direct gaze.

Copy-ready tested prompt
Task: Generate a professional headshot with a neutral background. Subject: a fully fictional adult East Asian woman aerospace engineer in her mid-30s, short textured black hair, warm brown eyes, a charcoal crew-neck jacket, and small matte silver stud earrings, with no resemblance to any real person. Context: a polished team biography portrait for a fictional research studio website. Style: photorealistic contemporary headshot photography with natural pores, fine facial texture, realistic fabric, and restrained retouching. Composition: chest-up at eye level, subject centered with comfortable shoulder room against a seamless warm gray background, 85 mm portrait perspective, direct calm gaze, landscape crop. Lighting: large soft key light slightly camera-left, gentle fill, subtle hair separation, neutral white balance, and soft natural catchlights. Constraints: preserve believable facial asymmetry and skin texture; no beauty-filter skin, logos, readable text, jewelry glare, extra people, watermark, border, or resemblance to a celebrity. Output: one clean 1536×1024 landscape professional portrait suitable as a verified GPT Image 2 prompt result.
Adjustable parameters
Background warmth, jacket color, lens perspective, crop, expression
Failure correction
If the face looks over-retouched, add ‘retain pores, fine lines, facial asymmetry, and natural under-eye texture; no beauty filter.’
Review focus
skin texture, eye sharpness, shoulder room, catchlights, and neutral color
Model
GPT Image 2
Portrait · GPT Image 2 · Prompt 02

Environmental Founder Portrait

For founder stories that need both a credible person and enough workplace context to explain the work.

GPT Image 2 portrait prompt example with a fictional founder in a materials workshop
Environmental Founder Portrait | Best for: For founder stories that need both a credible person and enough workplace context to explain the work. | Prompt ENP04-P02 | Check: deep-skin exposure, workshop hierarchy, hand anatomy, material realism, and brand safety | If it fails: If the setting overwhelms the face, add ‘reduce background contrast and detail; keep the subject as the only sharp, high-contrast focal point.’
Expected result

A confident editorial founder portrait where the workshop reads clearly without competing with the subject.

Copy-ready tested prompt
Task: Create an environmental founder portrait. Subject: a fully fictional adult Black woman renewable-materials founder in her mid-40s, deep brown skin, close-cropped natural hair, a moss-green work jacket over an ivory shirt, and no resemblance to any real person. Context: an editorial profile for a fictional circular-design company inside a small biocomposite workshop. Style: photorealistic business editorial photography with real skin texture, practical clothing, tactile materials, and understated color grading. Composition: waist-up three-quarter portrait at eye level, subject slightly right of center beside a workbench with unbranded fiber samples and abstract material tiles softly visible behind her, 50 mm perspective, landscape frame. Lighting: broad skylight from camera-left balanced by a warm practical task lamp and gentle shadow detail. Constraints: keep the workshop credible and secondary; no logos, readable labels, branded products, extra people, plastic skin, exaggerated machinery, watermark, or celebrity resemblance. Output: one polished 1536×1024 landscape environmental portrait suitable as a verified GPT Image 2 prompt result.
Adjustable parameters
Occupation, workshop materials, jacket color, camera distance, practical light
Failure correction
If the setting overwhelms the face, add ‘reduce background contrast and detail; keep the subject as the only sharp, high-contrast focal point.’
Review focus
deep-skin exposure, workshop hierarchy, hand anatomy, material realism, and brand safety
Model
GPT Image 2
Portrait · GPT Image 2 · Prompt 03

Creator-in-Studio Portrait

For magazine profiles and portfolio introductions that should show a maker inside a believable practice.

GPT Image 2 editorial portrait prompt result with a fictional ceramic sculptor in a studio
Creator-in-Studio Portrait | Best for: For magazine profiles and portfolio introductions that should show a maker inside a believable practice. | Prompt ENP04-P03 | Check: hand placement, clay residue, shelf depth, beard texture, and realistic studio scale | If it fails: If the studio feels staged, add ‘use a few functional, imperfect tools and works-in-progress; avoid decorative clutter and showroom styling.’
Expected result

A tactile, documentary-style studio portrait with natural posture, credible craft materials, and quiet visual depth.

Copy-ready tested prompt
Task: Generate a creator-in-studio portrait. Subject: a fully fictional adult Latino man ceramic sculptor in his late 30s, medium tan skin, dark wavy hair, a short natural beard, a faded indigo overshirt, and a clay-marked canvas apron, with no resemblance to any real person. Context: a magazine profile about an independent fictional craft practice. Style: photorealistic documentary-editorial portrait with honest skin texture, powdery clay residue, worn fabric, and a quiet handcrafted atmosphere. Composition: waist-up at eye level, subject leaning lightly against a clean worktable, three-quarter gaze toward camera, shelves of abstract unbranded ceramic forms softly out of focus, 50 mm perspective, landscape crop. Lighting: soft north-window light from camera-right, subtle bounce fill, muted warm highlights, and natural shadow falloff. Constraints: anatomically natural hands, believable studio scale, no logos, readable labels, extra people, luxury gloss, beauty retouching, watermark, or resemblance to a known artist. Output: one detailed 1536×1024 landscape creator studio portrait for a verified GPT Image 2 example.
Adjustable parameters
Creative discipline, apron wear, shelf objects, window direction, camera height
Failure correction
If the studio feels staged, add ‘use a few functional, imperfect tools and works-in-progress; avoid decorative clutter and showroom styling.’
Review focus
hand placement, clay residue, shelf depth, beard texture, and realistic studio scale
Model
GPT Image 2
Portrait · GPT Image 2 · Prompt 04

Natural Window-Light Portrait

For author, editor, consultant, or educator profiles that call for a quiet and approachable mood.

GPT Image 2 portrait lighting prompt result using natural window light
Natural Window-Light Portrait | Best for: For author, editor, consultant, or educator profiles that call for a quiet and approachable mood. | Prompt ENP04-P04 | Check: highlight rolloff, shadow detail, hair edges, linen texture, and background text safety | If it fails: If one side of the face goes muddy, add ‘place a white wall or reflector opposite the window and retain detail in the shadow cheek.’
Expected result

A softly modeled seated portrait with preserved shadow detail, gentle catchlights, and a natural reading-room atmosphere.

Copy-ready tested prompt
Task: Produce a natural window-light portrait. Subject: a fully fictional adult South Asian woman literary editor in her early 40s, warm medium-brown skin, shoulder-length softly waved black hair, subtle smile lines, and a pale blue linen shirt, with no resemblance to any real person. Context: an intimate author-profile image for a fictional independent journal. Style: photorealistic quiet editorial photography with realistic pores, linen weave, gentle tonal range, and no glamour retouching. Composition: seated chest-up three-quarter portrait beside a large window in a simple reading nook, body angled toward the light and eyes toward camera, blank book spines and one small plant softly blurred in the background, 85 mm perspective, landscape frame. Lighting: indirect overcast window light, white-wall bounce, soft catchlights, and preserved shadow detail without artificial rim light. Constraints: no logos, readable book titles, extra people, plastic skin, blown highlights, exaggerated bokeh, watermark, or celebrity resemblance. Output: one calm 1536×1024 landscape window-light portrait suitable as a tested GPT Image 2 prompt result.
Adjustable parameters
Window size, weather, seating angle, shirt fabric, background distance
Failure correction
If one side of the face goes muddy, add ‘place a white wall or reflector opposite the window and retain detail in the shadow cheek.’
Review focus
highlight rolloff, shadow detail, hair edges, linen texture, and background text safety
Model
GPT Image 2
Portrait · GPT Image 2 · Prompt 05

Outdoor Golden-Hour Portrait

For location profiles and travel editorials that need warmth without sacrificing accurate skin color.

GPT Image 2 realistic portrait prompt result outdoors at golden hour
Outdoor Golden-Hour Portrait | Best for: For location profiles and travel editorials that need warmth without sacrificing accurate skin color. | Prompt ENP04-P05 | Check: skin neutrality, rim separation, coastline simplicity, shirt texture, and flare control | If it fails: If skin turns orange, add ‘neutralize facial white balance and confine the warm color to rim light and background highlights.’
Expected result

A balanced environmental portrait with a fine sunset rim, neutral facial color, and an uncluttered landscape.

Copy-ready tested prompt
Task: Create an outdoor golden-hour portrait. Subject: a fully fictional adult Middle Eastern man ecological trail guide in his late 30s, olive-brown skin, short curly dark hair, neat stubble, and an unbranded sand-colored field shirt, with no resemblance to any real person. Context: a responsible travel editorial set in a fictional coastal grassland. Style: photorealistic location portrait photography with natural skin texture, wind-touched fabric, restrained cinematic color, and authentic outdoor detail. Composition: waist-up at eye level, subject placed on the left third and looking calmly toward camera, low grasses and a distant soft coastline behind him, 70 mm perspective, landscape crop with breathing room. Lighting: sun just above the horizon behind camera-right creating a fine rim on hair and shoulders, soft reflector fill on the face, warm highlights with neutral skin tones. Constraints: avoid orange skin, crushed shadows, lens-flare obscuring the face, logos, signage, tourists, landmarks, watermark, or resemblance to a public figure. Output: one balanced 1536×1024 landscape golden-hour portrait for a verified GPT Image 2 example.
Adjustable parameters
Sun position, reflector strength, field type, lens length, horizon height
Failure correction
If skin turns orange, add ‘neutralize facial white balance and confine the warm color to rim light and background highlights.’
Review focus
skin neutrality, rim separation, coastline simplicity, shirt texture, and flare control
Model
GPT Image 2
Portrait · GPT Image 2 · Prompt 06

Black-and-White Documentary Portrait

For community, nonprofit, and oral-history stories where expression and tonal structure matter more than color.

GPT Image 2 portrait prompt result in black-and-white documentary style
Black-and-White Documentary Portrait | Best for: For community, nonprofit, and oral-history stories where expression and tonal structure matter more than color. | Prompt ENP04-P06 | Check: midtones, wrinkle texture, microphone branding, note legibility, and eye detail | If it fails: If the blacks crush facial detail, add ‘lift the lower midtones while preserving a black point; keep both eyes and the shadow cheek readable.’
Expected result

An expressive monochrome portrait with open midtones, visible age detail, and a credible recording-room context.

Copy-ready tested prompt
Task: Create a black-and-white documentary portrait. Subject: a fully fictional adult East Asian man community radio volunteer in his early 60s, light-medium skin, silver hair brushed back, a weathered denim work shirt, and natural age lines, with no resemblance to any real person. Context: a nonprofit editorial story inside a fictional neighborhood recording room. Style: photorealistic monochrome documentary photography with rich midtones, subtle grain, honest skin texture, and restrained contrast. Composition: chest-up candid portrait at eye level, subject turned slightly away from an unbranded tabletop microphone and looking back toward camera, simple acoustic panels and blank pinned notes softly out of focus, 50 mm perspective, landscape crop. Lighting: one broad side window with a small white bounce, luminous eyes, detailed shadows, and no artificial glamour light. Constraints: monochrome only; preserve wrinkles and facial asymmetry; no logos, readable text, branded audio gear, extra people, crushed blacks, clipped whites, watermark, or public-figure resemblance. Output: one expressive 1536×1024 landscape black-and-white documentary portrait for a verified GPT Image 2 result.
Adjustable parameters
Grain strength, contrast curve, microphone position, crop, gaze direction
Failure correction
If the blacks crush facial detail, add ‘lift the lower midtones while preserving a black point; keep both eyes and the shadow cheek readable.’
Review focus
midtones, wrinkle texture, microphone branding, note legibility, and eye detail
Model
GPT Image 2
Portrait · GPT Image 2 · Prompt 07

Low-Key Cinematic Portrait

For dramatic program profiles, fictional production stills, and moody editorial introductions.

GPT Image 2 cinematic portrait prompt result with low-key lighting
Low-Key Cinematic Portrait | Best for: For dramatic program profiles, fictional production stills, and moody editorial introductions. | Prompt ENP04-P07 | Check: both-eye visibility, shadow color, skin tone, curtain detail, and edge-light restraint | If it fails: If an eye disappears, add ‘raise cool fill only on the shadow eye and cheek; keep the background dark but not featureless black.’
Expected result

A dark cinematic portrait whose face remains readable, dimensional, and naturally textured.

Copy-ready tested prompt
Task: Produce a low-key cinematic portrait. Subject: a fully fictional adult Southeast Asian man stage carpenter in his early 40s, medium-brown skin, close-cropped hair, a subtle mustache, and a dark burgundy work shirt, with no resemblance to any real person. Context: a dramatic profile image for a fictional independent theater program. Style: photorealistic cinematic portrait photography with natural pores, believable fabric, deep but detailed shadows, and restrained film color. Composition: seated chest-up three-quarter portrait in a simple black-box theater, subject centered slightly right, one wooden stage flat and soft curtain folds barely visible behind him, 85 mm perspective, landscape frame. Lighting: narrow warm key from camera-left at 45 degrees, very low cool fill, gentle edge light on one shoulder, catchlights retained, background falling to charcoal rather than pure black. Constraints: keep both eyes readable and skin tone accurate; no logos, stage titles, props suggesting a known production, extra people, smoke obscuring the face, plastic skin, watermark, or celebrity resemblance. Output: one polished 1536×1024 landscape low-key cinematic portrait for a tested GPT Image 2 example.
Adjustable parameters
Key width, fill ratio, background density, shirt color, edge-light strength
Failure correction
If an eye disappears, add ‘raise cool fill only on the shadow eye and cheek; keep the background dark but not featureless black.’
Review focus
both-eye visibility, shadow color, skin tone, curtain detail, and edge-light restraint
Model
GPT Image 2
Portrait · GPT Image 2 · Prompt 08

Colored-Gel Creative Portrait

For original music, arts, and creator profiles that need controlled saturated color without losing facial detail.

GPT Image 2 creative portrait prompt example with controlled colored-gel lighting
Colored-Gel Creative Portrait | Best for: For original music, arts, and creator profiles that need controlled saturated color without losing facial detail. | Prompt ENP04-P08 | Check: deep-skin exposure, eye clarity, color clipping, jacket detail, and background cleanliness | If it fails: If the gels contaminate all skin color, add ‘increase neutral frontal fill and reserve colored light for the sides and background.’
Expected result

A vivid studio portrait with clean magenta-cyan separation and accurate exposure across deep skin.

Copy-ready tested prompt
Task: Generate a creative colored-gel portrait. Subject: a fully fictional adult Black nonbinary electronic musician in their late 20s, deep brown skin, a short geometric natural haircut, a cobalt structured jacket over a plain black top, and no resemblance to any real person. Context: an original album-profile visual for a fictional independent music project with no branding. Style: photorealistic contemporary studio portrait with realistic pores, crisp fabric texture, saturated but controlled color, and editorial polish. Composition: shoulders-up at eye level, subject facing camera with a calm expression, centered against a dark neutral seamless background, 85 mm perspective, landscape crop with clean negative space. Lighting: magenta gel key from camera-left, cyan gel rim from camera-right, soft neutral frontal fill to preserve true facial detail, distinct color separation without clipping. Constraints: accurate deep-skin exposure, both eyes clearly visible, no logos, instruments, readable text, neon signs, extra people, oily highlights, oversaturation, watermark, or resemblance to a celebrity musician. Output: one vivid 1536×1024 landscape colored-gel portrait suitable as a verified GPT Image 2 prompt result.
Adjustable parameters
Gel hues, saturation, neutral fill, jacket color, negative space
Failure correction
If the gels contaminate all skin color, add ‘increase neutral frontal fill and reserve colored light for the sides and background.’
Review focus
deep-skin exposure, eye clarity, color clipping, jacket detail, and background cleanliness
Model
GPT Image 2
Portrait · GPT Image 2 · Prompt 09

Realistic Skin-Texture Close-Up

For editorial close-ups that must look human rather than airbrushed or cosmetically reconstructed.

GPT Image 2 realistic portrait prompt result preserving natural skin texture
Realistic Skin-Texture Close-Up | Best for: For editorial close-ups that must look human rather than airbrushed or cosmetically reconstructed. | Prompt ENP04-P09 | Check: freckles, peach fuzz, lashes, under-eye texture, and sharpening artifacts | If it fails: If texture becomes crunchy, add ‘resolve natural pores without high-radius sharpening, HDR halos, or exaggerated microcontrast.’
Expected result

A tight portrait with visible freckles, pores, fine hairs, and smooth highlight rolloff while keeping the eyes crisp.

Copy-ready tested prompt
Task: Create a close-up portrait that preserves realistic skin texture. Subject: a fully fictional adult white woman orchard researcher in her late 30s, fair skin with natural freckles and mild sun texture, green-gray eyes, loosely tied auburn hair, and a plain forest-green shirt, with no resemblance to any real person. Context: an educational editorial portrait about field science for a fictional publication. Style: highly photorealistic close portrait photography with visible pores, peach fuzz, fine lines, individual lashes, and restrained natural color. Composition: tight head-and-shoulders crop at eye level, face turned ten degrees toward camera, eyes sharply focused, softly blurred neutral field shelter in the background, 100 mm perspective, landscape frame. Lighting: large diffused open-shade light, small white bounce below eye line, gentle catchlights, and smooth highlight rolloff. Constraints: retain freckles, pores, under-eye texture, facial asymmetry, and flyaway hairs; no skin smoothing, pore removal, beauty makeup, excessive sharpening, logos, text, extra people, watermark, or celebrity resemblance. Output: one detailed 1536×1024 landscape close-up portrait demonstrating natural skin texture for a verified GPT Image 2 result.
Adjustable parameters
Crop tightness, pore detail, background blur, reflector height, focal length
Failure correction
If texture becomes crunchy, add ‘resolve natural pores without high-radius sharpening, HDR halos, or exaggerated microcontrast.’
Review focus
freckles, peach fuzz, lashes, under-eye texture, and sharpening artifacts
Model
GPT Image 2
Portrait · GPT Image 2 · Prompt 10

Respectful Mature-Adult Portrait

For community and expert profiles that should preserve age, dignity, and natural expression.

GPT Image 2 realistic portrait prompt result of a fictional mature adult
Respectful Mature-Adult Portrait | Best for: For community and expert profiles that should preserve age, dignity, and natural expression. | Prompt ENP04-P10 | Check: age fidelity, hand anatomy, silver coils, skin highlights, and stereotype avoidance | If it fails: If the subject is de-aged, add ‘preserve natural wrinkles, skin laxity, gray hair, neck detail, and apparent age without glamour retouching.’
Expected result

A warm greenhouse portrait with authentic wrinkles, silver hair detail, relaxed posture, and accurate deep-skin highlights.

Copy-ready tested prompt
Task: Generate a natural portrait of a mature adult. Subject: a fully fictional adult Black woman urban botanist in her late 60s, deep warm-brown skin, short silver coils, expressive smile lines, practical round brass earrings, and a muted teal linen overshirt, with no resemblance to any real person. Context: a respectful community-profile portrait inside a fictional teaching greenhouse. Style: photorealistic human-centered editorial photography with authentic age texture, soft fabric detail, natural posture, and calm color. Composition: waist-up at eye level, subject standing beside a plain potting bench, turned slightly toward camera with a composed half-smile, leafy forms softly blurred behind her, 70 mm perspective, landscape crop. Lighting: broad greenhouse daylight filtered through translucent panels, gentle negative fill for facial shape, soft eye catchlights, and accurate highlight retention on deep skin. Constraints: preserve wrinkles, gray hair texture, natural neck and hand detail; no age reduction, beauty-filter skin, stereotypes, logos, plant labels, extra people, watermark, or resemblance to a public figure. Output: one dignified 1536×1024 landscape mature portrait suitable as a tested GPT Image 2 prompt result.
Adjustable parameters
Age range, environment, wardrobe color, expression, negative fill
Failure correction
If the subject is de-aged, add ‘preserve natural wrinkles, skin laxity, gray hair, neck detail, and apparent age without glamour retouching.’
Review focus
age fidelity, hand anatomy, silver coils, skin highlights, and stereotype avoidance
Model
GPT Image 2
Portrait · GPT Image 2 · Prompt 11

Anti-Glare Glasses Portrait

For professional headshots where clear eyes and undistorted lenses are essential.

GPT Image 2 professional headshot prompt result with reflection-free glasses
Anti-Glare Glasses Portrait | Best for: For professional headshots where clear eyes and undistorted lenses are essential. | Prompt ENP04-P11 | Check: lens reflections, eye visibility, frame symmetry, facial texture, and backdrop smoothness | If it fails: If reflections remain, add ‘move both sources higher and farther to the sides, feather them across the face, and keep the lens axis outside their reflection path.’
Expected result

A clean staff portrait with both eyes fully visible through matte frames and no bright geometric reflections.

Copy-ready tested prompt
Task: Produce a professional portrait of a person wearing glasses without distracting reflections. Subject: a fully fictional adult East Asian man accessibility researcher in his early 30s, light-medium skin, straight dark hair with a soft side part, thin matte charcoal rectangular glasses, and a navy knit polo, with no resemblance to any real person. Context: a staff headshot for a fictional inclusive-technology laboratory. Style: photorealistic modern professional portrait with realistic pores, crisp knit texture, subtle retouching, and neutral color. Composition: chest-up at eye level, subject angled slightly camera-left while looking directly at the lens, centered against a soft pale-slate backdrop, 85 mm perspective, landscape crop. Lighting: two large feathered softboxes positioned above and far to each side so both lenses remain clear, low frontal fill, natural catchlights, and neutral white balance. Constraints: eyes must be fully visible through both lenses; no ring-light circles, window glare, green reflections, lens distortion, logos, readable text, extra people, plastic skin, watermark, or celebrity resemblance. Output: one clean 1536×1024 landscape anti-glare glasses portrait for a verified GPT Image 2 example.
Adjustable parameters
Frame shape, softbox height, subject angle, backdrop tone, fill level
Failure correction
If reflections remain, add ‘move both sources higher and farther to the sides, feather them across the face, and keep the lens axis outside their reflection path.’
Review focus
lens reflections, eye visibility, frame symmetry, facial texture, and backdrop smoothness
Model
GPT Image 2
Portrait · GPT Image 2 · Prompt 12

Curl and Flyaway Detail Portrait

For creator profiles and beauty-neutral close portraits where natural hair edges must remain legible.

GPT Image 2 portrait prompt result preserving curls and natural flyaway hairs
Curl and Flyaway Detail Portrait | Best for: For creator profiles and beauty-neutral close portraits where natural hair edges must remain legible. | Prompt ENP04-P12 | Check: curl pattern, flyaway realism, edge halos, skin texture, and hair-frame completeness | If it fails: If curls melt into a painted mass, add ‘resolve grouped curl clumps and a few separate wisps; avoid uniform strand repetition and edge halos.’
Expected result

A crisp portrait with individually resolved curls, a few believable flyaways, and clean separation from a dark background.

Copy-ready tested prompt
Task: Create a high-detail portrait that resolves tight curls and fine flyaway hairs. Subject: a fully fictional adult mixed-heritage woman sound editor in her early 30s, medium warm-brown skin, dense shoulder-length dark 3C curls with a few natural flyaways, hazel-brown eyes, and a rust-colored textured sweater, with no resemblance to any real person. Context: a creative-industry profile for a fictional postproduction studio. Style: photorealistic editorial portrait photography with realistic pores, individually resolved curls, soft fabric texture, and natural color. Composition: shoulders-up at eye level, subject turned slightly camera-right and looking into the lens, hair fully inside the frame against a smooth deep-gray background, 85 mm perspective, landscape crop. Lighting: large soft frontal-side key, narrow controlled backlight to separate curl edges, gentle fill on the shadow cheek, and clean catchlights. Constraints: preserve curl pattern, stray wisps, hair volume, and natural edges without crunchy sharpening; no melted hair, painted strands, halo clipping, logos, readable text, extra people, plastic skin, watermark, or celebrity resemblance. Output: one crisp 1536×1024 landscape curl-detail portrait suitable as a verified GPT Image 2 result.
Adjustable parameters
Curl pattern, backlight width, hair volume, crop, sweater texture
Failure correction
If curls melt into a painted mass, add ‘resolve grouped curl clumps and a few separate wisps; avoid uniform strand repetition and edge halos.’
Review focus
curl pattern, flyaway realism, edge halos, skin texture, and hair-frame completeness
Model
GPT Image 2
Portrait · GPT Image 2 · Prompt 13

Accurate Deep-Skin Exposure

For workplace and executive portraits that require rich undertones, controlled highlights, and open shadows.

GPT Image 2 portrait lighting prompt result with accurate very deep skin exposure
Accurate Deep-Skin Exposure | Best for: For workplace and executive portraits that require rich undertones, controlled highlights, and open shadows. | Prompt ENP04-P13 | Check: undertones, forehead highlights, facial contour, tool branding, and orange-shirt separation | If it fails: If shadows look ashy, add ‘use neutral broad fill, preserve warm-neutral undertones, and avoid lifting shadows with gray desaturation.’
Expected result

A calibrated professional portrait with luminous very deep skin, neutral shadows, and realistic laboratory context.

Copy-ready tested prompt
Task: Produce a portrait with accurate exposure and color for very deep skin. Subject: a fully fictional adult woman precision-instrument technician in her late 30s, very deep neutral-brown skin, close-twisted black hair gathered into a low bun, dark brown eyes, and a simple burnt-orange collared shirt, with no resemblance to any real person. Context: an editorial workplace profile for a fictional calibration laboratory with all equipment unbranded. Style: photorealistic professional portrait photography with true-to-life undertones, visible pores, controlled highlights, and restrained neutral grading. Composition: chest-up at eye level, subject positioned slightly left of center with a calm direct gaze, softly blurred matte metal tools and a blank gray wall behind her, 70 mm perspective, landscape frame. Lighting: very large diffused key close to camera-left, broad neutral fill, subtle warm rim, calibrated white balance, and enough exposure to retain both highlight and shadow detail. Constraints: maintain rich natural skin color without ashy gray shadows, red oversaturation, clipped forehead highlights, or crushed facial contours; no logos, readable instrument markings, extra people, plastic skin, watermark, or public-figure resemblance. Output: one carefully exposed 1536×1024 landscape deep-skin portrait for a tested GPT Image 2 example.
Adjustable parameters
Key size, fill ratio, white balance, shirt color, background brightness
Failure correction
If shadows look ashy, add ‘use neutral broad fill, preserve warm-neutral undertones, and avoid lifting shadows with gray desaturation.’
Review focus
undertones, forehead highlights, facial contour, tool branding, and orange-shirt separation
Model
GPT Image 2
Portrait · GPT Image 2 · Prompt 14

Identity-Preserving Background Edit

For rights-cleared staff portrait updates where only the setting should change and the synthetic person must stay fixed.

Authorized synthetic input Authorized synthetic input portrait created for the GPT Image 2 background edit example
GPT Image 2 edited output GPT Image 2 image-editing portrait result preserving a fictional identity
Identity-Preserving Background Edit | Best for: For rights-cleared staff portrait updates where only the setting should change and the synthetic person must stay fixed. | Prompt ENP04-P14 | Check: face identity, jacket shape, hair boundary, lighting match, and absence of readable titles | If it fails: If identity drifts, add ‘mask and preserve the person pixel-for-pixel where possible; regenerate only the background and boundary transition.’
Expected result

A seamless background replacement that keeps facial geometry, pose, wardrobe, hair edges, and original portrait lighting intact.

Copy-ready tested prompt
Task: Replace only the background of the provided synthetic portrait while preserving the person’s identity. Subject: keep the fully fictional adult woman exhibition architect from the authorized GPT Image 2 source image exactly the same, including her medium-deep brown skin, almond-shaped dark eyes, short tapered natural hairstyle, matte brass stud earrings, slate-blue collarless jacket, ivory top, facial geometry, expression, and apparent age. Context: a rights-cleared identity-preservation test for a fictional museum staff profile; the input was generated specifically for this example and depicts no real person. Style: photorealistic professional editorial portrait with the source image’s natural pores, believable fabric, restrained retouching, and neutral color. Composition: retain the exact waist-up eye-level framing, centered pose, gaze, shoulder position, body proportions, and 70 mm perspective; change only the seamless gray background to a bright contemporary exhibition-library interior with pale wood shelving, blank book spines, and softly blurred unbranded display plinths. Lighting: preserve the original facial key, fill, hair light, catchlights, skin exposure, and jacket shadows; make the new background light direction and depth of field consistent with that portrait. Constraints: do not alter, beautify, redraw, resize, or reposition the person or clothing; preserve clean hair and jacket edges; no extra people, logos, readable titles, museum branding, new jewelry, plastic skin, halos, watermark, or celebrity resemblance. Output: one seamless 1536×1024 landscape background-replacement portrait that visibly preserves the authorized synthetic identity.
Adjustable parameters
Replacement setting, background blur, shelf density, ambient color, edge cleanup
Failure correction
If identity drifts, add ‘mask and preserve the person pixel-for-pixel where possible; regenerate only the background and boundary transition.’
Review focus
face identity, jacket shape, hair boundary, lighting match, and absence of readable titles
Model
GPT Image 2
Portrait · GPT Image 2 · Prompt 15

Three-Composition Contact Sheet

For planning headshot, chest-up, and environmental crops before a professional portrait campaign.

GPT Image 2 portrait prompt result comparing three consistent portrait compositions
Three-Composition Contact Sheet | Best for: For planning headshot, chest-up, and environmental crops before a professional portrait campaign. | Prompt ENP04-P15 | Check: face consistency, silver-streak position, blazer continuity, crop differences, and panel count | If it fails: If the person changes between panels, add ‘reuse one locked identity and one studio session; only camera distance and surrounding context may change.’
Expected result

Three visibly different crops of one stable fictional person, with matching identity, wardrobe, light, and color.

Copy-ready tested prompt
Task: Create a three-frame contact sheet showing the same fictional person in three portrait compositions. Subject: one fully fictional adult woman landscape architect in her mid-40s, medium olive skin, a chin-length wavy dark-brown bob with one silver streak at the left temple, hazel eyes, and the identical moss-green blazer over a cream crew-neck shirt in every frame, with no resemblance to any real person. Context: an educational photography planning sheet comparing crop choices for a fictional professional profile. Style: photorealistic editorial portrait photography with consistent identity, natural pores, realistic fabric, and restrained neutral grading. Composition: one 3:2 landscape canvas divided into exactly three equal vertical panels with subtle plain gutters: left panel tight headshot, center panel chest-up portrait, right panel waist-up environmental portrait beside an unbranded drafting table; same eye-level camera family and direct calm gaze in all panels. Lighting: identical large soft key from camera-left, gentle fill, subtle hair separation, and matching white balance across all three panels. Constraints: preserve exactly the same face, apparent age, bob shape, silver streak position, eye color, blazer, shirt, and skin tone across panels; no extra people, duplicate limbs, logos, readable drawings, labels, panel text, watermark, border, or celebrity resemblance. Output: one clean 1536×1024 landscape three-composition contact sheet with all three panels fully visible.
Adjustable parameters
Crop set, panel count, wardrobe, environment, gutter width
Failure correction
If the person changes between panels, add ‘reuse one locked identity and one studio session; only camera distance and surrounding context may change.’
Review focus
face consistency, silver-streak position, blazer continuity, crop differences, and panel count
Model
GPT Image 2

GPT Image 2 portrait prompt FAQ

What should a GPT Image 2 portrait prompt include?

Include the task, a fictional or authorized subject, the intended context, photographic style, crop and camera perspective, lighting design, explicit constraints, and output dimensions. Concrete decisions are more useful than a long list of flattering adjectives.

How do I get realistic skin instead of a beauty-filter look?

Ask to preserve pores, peach fuzz, fine lines, freckles, facial asymmetry, and under-eye texture. Request restrained retouching and natural highlight rolloff, then explicitly exclude pore removal, plastic skin, excessive sharpening, and glamour smoothing.

How can I prevent reflections in glasses?

Place large feathered sources higher and farther to each side so their reflection paths miss the lenses. Keep both eyes visible, use matte frames, avoid ring lights and bright windows in front of the subject, and inspect each lens independently.

What is the safest way to edit a portrait background?

Use only an image you own or are authorized to edit. Tell GPT Image 2 to change the background alone, freeze the person’s facial geometry, expression, pose, crop, clothing, age, skin exposure, and catchlights, and review hair and clothing edges for halos or redraws.

How do I expose very deep skin accurately?

Use a large diffused key, broad neutral fill, calibrated white balance, and controlled highlights. Ask for rich natural undertones with readable shadow contours, and exclude ashy gray lift, red oversaturation, crushed blacks, and clipped forehead highlights.

Can I copy these GPT Image 2 portrait prompts for commercial work?

The examples use fictional adults, generic settings, and an authorized synthetic edit input without brands, celebrities, protected characters, or living-artist style requests. Check your own substituted people, source-image rights, model terms, employer policies, releases, and local law before commercial use.