GPT-Image-2 프롬프트 모음|이미지 생성에 바로 쓰는 엄선 실전 예시와 결과 총정리 2026

gptimage2on 4 hours ago

이 글은 Twitter(X)에서 크리에이터들이 공유한 실제 GPT-Image-2 프롬프트를 모아 계속 업데이트하는 컬렉션입니다. 각 항목은 원본 게시물로 연결됩니다. 새로운 프롬프트가 발견될 때마다 목록을 추가하니 정기적으로 확인해 보세요.


데이터 시각화: 이중 축 차트

공유자: @maxescu

DOMAIN: "Crime, Safety & Justice"

HEADLINE: "Violent crime in the US has fallen 50% since 1993. Most Americans think it's at an all-time high."

Use Google Search to find: US violent crime rate trend 1990–today (FBI UCR/BJS NCVS data), percentage decline from 1993 peak to today, and Gallup polling data showing % of Americans who believe crime is increasing (consistently 60–70%+ even as crime falls). Cream #F7F3EC. Square.

HEADLINE: "Crime is down 50% since 1993. Most Americans think it's going up. Both true simultaneously."

SUBHEAD: "US VIOLENT CRIME RATE vs PUBLIC PERCEPTION · 1990–PRESENT · BJS/FBI + GALLUP"

CHART: Dual-axis diverging chart.

LINE A (electric blue): US violent crime rate per 100,000 — sharp peak 1991–1993, dramatic decline, near historic lows today.

LINE B (coral dashed): % of Americans saying crime is "up from last year" (Gallup) — stays high, barely reflects actual trend. The two lines diverge dramatically. Reality vs perception.

ANNOTATION: "The crime rate and the public's feeling about crime have almost no relationship."

SOURCE: "Bureau of Justice Statistics NCVS · FBI Uniform Crime Reports · Gallup Crime Poll"
GPT-Image-2 이중 축 차트: 미국 폭력 범죄율과 대중 인식 비교

초고해상도 도시 장면

공유자: @mark_k, 결과 테스트: @WolfRiccardo

A hyper-detailed photograph captured on a Canon EOS R5 DSLR camera, 24-105mm f/4L lens at 35mm, f/5.6 aperture, 1/250s shutter, ISO 400, golden-hour dusk lighting with soft lens flare and subtle chromatic aberration. Epic bustling Times Square-style metropolitan intersection teeming with chaotic urban life: over 150 diverse pedestrians of every ethnicity, age, and attire (business suits on phones, tourists with backpacks and maps, families pushing strollers, street dancers performing, elderly with canes, cyclists, joggers, vendors in aprons). Yellow taxis, electric buses, motorcycles, delivery scooters, and hovering drones weave through heavy traffic. Towering glass skyscrapers reflect the vibrant chaos; wet pavement from light rain glistens with perfect specular highlights and steam rising from manholes. Foreground newspaper kiosk overflows with papers headlined "BREAKING: GPT-Image-2 Redefines AI Art 2026" and magazines showing crisp covers. Multiple illuminated billboards and neon signs feature perfectly legible text: "Coca-Cola Zero", "Nike Air Max 2026", "Tesla Full Self-Driving", "Starbucks Reserve", "iPhone 17 Pro Max – Preorder Now", "NYC Marathon – Nov 2026", "Macy's 70% Off Flash Sale". A massive central 3D digital billboard prominently glows with bold futuristic lettering "GPT-Image-2". Shop windows display "OPEN 24/7 – Fresh Bagels", phone screens show social feeds, license plates read "AI-2026" and "PRO-MODE", traffic signs say "No Honking" and "Pedestrian Crossing". Pigeons scatter, hot-dog carts emit smoke, distant helicopter hovers.
GPT-Image-2로 만든, 광고판 글자까지 선명한 초고해상도 타임스스퀘어풍 도시 장면WolfRiccardo가 테스트한 GPT-Image-2 초고해상도 도시 장면 결과

사실적인 인물 사진: 섬세한 플래시 촬영

공유자: @BubbleBrain

CCD compact camera direct on-camera flash, full-body portrait, eye-level straight-on shot, gorgeous Korean idol woman, early 20s, 9-head body proportion, standing in confident S-curve pose with weight on back leg, front leg slightly forward and bent at knee, elongating leg line, one hand forming finger-gun gesture pressed against temple, head tilted, sharp cold gaze directly into lens, lips slightly parted, expression reads effortless danger — sexy and unbothered, long jet-black pin-straight hair reaching lower back, silky texture with fine strand separation under flash, no bangs, hair tucked behind one ear revealing jawline, cold-toned porcelain skin with authentic texture — visible fine pores on cheeks and nose, soft peach fuzz on jawline and cheekbones catching light, subtle subsurface scattering at ear edges and nose tip, natural skin micro-texture with faint expression lines, matte healthy glow with zero oiliness or plastic sheen, cool-white undertone, no yellow cast, no tan, refined Korean makeup — clean elongated cat eyeliner, individual lash strands visible, groomed natural brows, soft nude-pink matte lip, subtle highlighter on nose bridge and cheekbone peaks, small beauty mark under left eye, black thin-strap deep-V camisole top, fitted, fabric clinging to torso sculpting slim waist and lifted bust, black high-waisted bodycon mini skirt ending mid-thigh, fabric hugging hip curves with visible tension at seams, bare long toned legs with realistic skin texture — faint knee definition, subtle muscle tone, natural skin color variation, black pointed-toe stiletto heels 10cm, direct on-camera flash lighting, hard light quality, properly exposed — skin detail fully preserved with no blowout, characteristic CCD flash contrast: subject crisp and bright, background falling off to deep shadow behind her, short hard shadow outline cast on wall directly behind, slight natural vignette at frame corners, CCD color rendition — punchy midtone contrast, slightly cool white balance from flash mixing with warm ambient, ultra-realistic skin detail under harsh direct light, every pore and fine hair strand rendered, fabric weave texture visible on black garments, metallic sheen on stiletto heel catching flash, hair strands individually lit with scattered micro-highlights, sharp critical focus on eyes and face, 8K UHD, RAW photograph, masterpiece, best quality, photorealistic, zero stylization, real human being
GPT-Image-2로 만든 한국 아이돌풍 사실적 CCD 플래시 인물 사진

패션 & 라이프스타일: Pinterest 스타일 코디

공유자: @WolfRiccardo

Ref: outfit/combo - Woman. Pinterest vibe - cool pose.
The woman should be wearing only one sleeve of her jacket, and she should be holding the bag with that same arm.
GPT-Image-2 Pinterest 스타일 패션 코디 프롬프트 결과GPT-Image-2 Pinterest 스타일 패션 코디 두 번째 결과

메이크업 & 뷰티: 2분할 비교

공유자: @WolfRiccardo

These tones should be present. The woman should have applied eyeshadow and eye makeup in these tones.

It should be a 2:3 aspect ratio. It should be a two-panel image. One portrait, one close-up of the face/eyes (eyes closed, focus on makeup).

Should be realistic RAW quality, in a home environment.

It should have realistic skin textures and tones. Realistic facial structure and detail. She should have finished her makeup with gloss. It should be an pinterest-style image. It should give off a social media/Pinterest vibe.
GPT-Image-2 메이크업 뷰티 2분할 인물 사진 결과GPT-Image-2 메이크업 뷰티 눈 클로즈업 결과

영화 촬영 분석 주석

공유자: @VictorInFocus

GPT-Image-2를 사용해 『블레이드 러너』 영화 스틸을 분석하고, 조명·구도·렌즈·블로킹 라벨을 이미지 위에 정확하게 표시한 사례입니다.

Analyze this film still and annotate the shot. Overlay:
- lighting direction and quality
- composition rules used
- lens characteristics (focal length, depth of field)
- blocking and actor positioning

Place labels accurately on the image. Clean typography, cinematic analysis style.
GPT-Image-2로 만든 블레이드 러너 스틸컷 영화 촬영 분석 주석GPT-Image-2로 만든 조명과 구도 라벨이 포함된 블레이드 러너 샷 분석

빈티지 여행 포스터 일러스트

공유자: @WolfRiccardo

Modern pencil illustration of Vintage travel poster illustration of the Amalfi Coast, Italy, panoramic coastal cliff road scene, classic 1960s white car driving along a curved seaside road, deep blue Mediterranean sea with small sailboats, colorful pastel hillside village, bright blue sky with soft clouds, lemon tree branches with vibrant yellow lemons framing the foreground, warm summer sunlight, bold vibrant colors, retro 1950s travel poster style, cinematic composition, high detail, screen print texture, graphic illustration. Hand-drawn style, illustration with loose strokes and defined contours. High-contrast color palette, maintaining chromatic harmony between background and elements. Contemporary and decorative aesthetic.
GPT-Image-2로 만든 아말피 해안 빈티지 여행 포스터풍 연필 일러스트

간단한 프롬프트로 만든 멋진 게임 커버

공유자: @mirochill

Cover switch game : Discord Mod Simulator
GPT-Image-2로 만든 Discord Mod Simulator Nintendo Switch 게임 커버

세계 지식: 젤다 + 실제 열차

공유자: @marmaduke091

In the game Zelda TOTK, Link is in an E531 series train made by him
GPT-Image-2로 만든 실제 JR 동일본 E531계 열차 안의 젤다 TOTK 링크

직광 플래시를 활용한 아날로그 에디토리얼 인물 사진

공유자: @BubbleBrain

Analog 35mm film photography, high-fashion editorial portrait with intense on-camera direct flash (CCD hard flash style), intimate low-angle POV shot from slightly below (仰拍, around +15° upward angle), emphasizing long legs and elegant hourglass body proportions, 4:3 vertical format, 3/4 body shot, early 20s Korean idol-type young woman with delicate refined features, slightly upturned almond-shaped fox eyes, small sharp jawline, flawless porcelain skin with cool ivory undertone and visible specular highlights from direct flash, soft coral blush, natural dusty rose lips slightly parted, long naturally wavy dark brown hair with loose strands framing the face, wearing a slightly tilted black police peaked cap with silver badge.

She wears a fitted, stylish policewoman-inspired outfit: black short-sleeved collared crop top with shiny silver buttons and epaulets, a shiny silver police badge pinned on the chest, a thin black necktie loosely hanging at the front, a wide black duty belt with a silver buckle and handcuffs at the side, and a short black pleated mini skirt. Black thigh-high stockings and white low-top canvas sneakers with loose lacing add a casual contrast.

Pose: a subtle S-curve stance with weight shifted to one leg, hips angled slightly toward the camera to enhance silhouette and leg length. One hand lightly touches the cap brim or rests near the waist, while the other hand is placed at the hip or near the belt. Head tilted slightly down with a confident, self-assured upward gaze through lashes directed at the viewer, expression poised, cool, and subtly playful rather than overtly seductive.

Harsh on-camera direct flash from front/slightly above creates strong catchlights, sharp shadows behind the subject, crisp highlights on skin, hair, badge, belt buckle, and fabric textures. Cold flash quality mixed with warm ambient window light, visible flash fall-off, Kodak Portra 400 film simulation pushed for strong analog character, warm peachy skin tones with flash highlights, subtle cyan shadows, prominent heavy analog film grain (strength 0.65, larger grain size), noticeable film halation around bright areas, strong natural vignette, slight underexposure for a raw intimate mood.

Ultra-realistic photorealistic film photography, edgy magazine editorial aesthetic, visible fabric texture and reflective metal details, delicate hair strands glowing from flash, natural skin texture with heavy film grain throughout, no plastic skin, no digital sharpness, no airbrushing, no watermark, no text, authentic 35mm film look with strong grain and flash artifacts, analog color grading.
GPT-Image-2로 만든 아날로그 35mm 직광 플래시 에디토리얼 인물 사진

역사 속 인물의 소셜 미디어: 오다 노부나가

공유자: @itnavi2022

織田信長のX投稿ページ(本能寺の変直前)を作成してください。

의미:

혼노지의 변 직전 오다 노부나가의 X 게시물 페이지를 만든다.
GPT-Image-2로 만든 혼노지의 변 직전 오다 노부나가의 역사풍 X 게시물 페이지

역사 속 인물의 소셜 미디어: 사카모토 료마

공유자: @itnavi2022

日本の将来について熱く語る坂本龍馬のX投稿ページを作成して。

의미:

사카모토 료마가 일본의 미래에 대해 열정적으로 이야기하는 X 게시물 페이지를 만든다.
GPT-Image-2로 만든 사카모토 료마가 일본의 미래를 이야기하는 역사풍 X 게시물 페이지

콘셉트 제품 디자인: 심리스 재킷

공유자: @SRKDAN

A jacket with no zip, no buttons, no fastening — closes perfectly.
GPT-Image-2로 만든 지퍼, 단추, 잠금장치가 없는 심리스 재킷 콘셉트

콘셉트 제품 디자인: 크라운 없는 시계

공유자: @SRKDAN

A watch with no crown, no buttons — a sealed case that sets itself.
GPT-Image-2로 만든 밀폐 케이스와 자동 설정 기능의 크라운 없는 시계 콘셉트

콘셉트 제품 디자인: 심리스 지갑

공유자: @SRKDAN

A leather wallet with no stitching, no clasp, no visible opening.
GPT-Image-2로 만든 바느질, 잠금장치, 보이는 개구부가 없는 심리스 가죽 지갑 콘셉트

이 목록은 정기적으로 업데이트됩니다. 추가하고 싶은 프롬프트가 있다면 Twitter에 공유하고 GPT-Image-2 태그를 달아 주세요.