These GLM-Image prompt examples are organized by use case: portrait, landscape, product, poster, infographic, and bilingual signage. Every prompt is ready to paste into the generator as-is, and every one comes with a one-line note on why it works — because the goal is not a library you raid blindly, it is a set of patterns you can rebuild for your own subjects. Adapt them one variable at a time; wholesale rewrites reset everything the prompt got right.
All of them follow the structure covered in how to write AI image prompts: subject, then style, then concrete details. If a prompt here confuses you, that article explains the grammar behind it.
<!--IMG: A grid of six generated images spanning portrait, landscape, product, poster, infographic, and bilingual signage styles-->
Portrait prompt examples
Professional headshot of a software engineer in her 30s, short
black hair, slight knowing smile, direct eye contact, navy blazer
over gray tee, studio softbox lighting with subtle rim light,
plain warm-gray background, 85mm lens, sharp focus on the eyes
Why it works: identity, expression, wardrobe, light direction, and background are all decided — nothing left to average.
Weathered fisherman in his 60s tying rope on a wooden dock,
overcast North Atlantic light, salt-and-pepper beard, yellow
slicker, documentary photography, 50mm lens, muted color palette
Why it works: character comes from specifics (rope, slicker, overcast light), not from the word "character."
Candid portrait of a street vendor laughing behind a fruit stall,
late afternoon sun raking across the produce, shallow depth of
field, blurred market crowd behind, photojournalism style
Why it works: the light does the storytelling — one direction, one quality, stated plainly.
Fashion editorial portrait, model in a structured crimson dress
against a bare concrete wall, single hard spotlight from upper
right, deep shadows, glossy magazine aesthetic, medium format look
Why it works: one light source, one color anchor, one mood — restraint reads as expensive.
Landscape prompt examples
Snow-capped mountain ridge at sunrise, alpine lake mirroring the
peak, low fog in the valley, evergreen foreground silhouette,
landscape photography, 24mm wide angle, layered blues and golds
Why it works: foreground, midground, reflection, and light time are each named — composition builds itself.
Autumn birch forest with a narrow dirt path, morning mist between
white trunks, fallen leaves carpeting the ground, soft diffused
light, natural color palette, vertical composition
Why it works: "vertical composition" and "mist between trunks" are the two decisions that make it a photo, not a screensaver.
Rice terraces at dusk reflecting a violet sky, single farmhouse
with one lit window, layered green and gold bands descending the
hillsides, high vantage point, serene atmosphere, photorealistic
Why it works: the lit window is a focal point; without one, wide landscapes drift.
Storm approaching over coastal cliffs, massive gray clouds with
one break of golden light hitting the water, white spray against
dark rock, dramatic wide-angle landscape photography, high contrast
Why it works: the "one break of light" instruction gives the model a hero moment to render.
Product photography prompt examples
Matte-black wireless headphones floating above a walnut surface,
single softbox from upper left creating one clean highlight on the
ear cup, deep charcoal gradient background, three-quarter angle,
commercial product photography, sharp focus
Why it works: material, light direction, surface, and angle are the four legs a product shot stands on.
Clear glass perfume bottle with amber liquid on wet black slate,
backlit to make the liquid glow, water droplets on the surface,
single jasmine flower out of focus in the foreground, luxury
fragrance advertising style
Why it works: backlighting plus one out-of-focus prop is a classic fragrance formula, stated explicitly.
Craft coffee bag in kraft paper with a minimal cream label,
standing on light oak, morning window light from the right casting
a soft long shadow, warm tones, lifestyle e-commerce photography
Why it works: the shadow is invited, not tolerated — "soft long shadow" turns light into composition.
White running shoe mid-stride frozen against a seamless studio
background, dramatic side lighting emphasizing the knit texture,
subtle motion blur on the background, sports catalog aesthetic
Why it works: texture-forward lighting suits knit fabric; "frozen mid-stride" supplies energy without chaos.
Poster prompt examples (with text)
Music festival poster, bold sans-serif headline "SOUND WAVES"
centered at top, purple-to-black gradient background, single white
line-art wave motif across the middle, small date line "JULY 18"
at the bottom edge, flat modern graphic design
Why it works: exact copy in quotes, stated placement, one motif — poster discipline in one prompt.
Minimalist bookstore poster, serif title "SLOW READING DAY"
stacked in the upper third, cream background, delicate line
illustration of an open book with a plant growing from it,
generous negative space, muted greens and browns
Why it works: "generous negative space" is an instruction, and short text renders reliably.
Exhibition poster for a photography retrospective, duotone portrait
occupying the lower two-thirds, large condensed sans-serif title
"CITY LIGHTS" running down the left edge, thin white rule lines,
gallery-grade typography
Why it works: the two-thirds/one-third split and edge-run title give GLM-Image a real layout to hold.
Vintage travel poster of a coastal town at golden hour, warm
limited palette, stylized sun and sailboat, bold retro lettering
"CAPE BAY" in an arc at the top, screen-print texture
Why it works: naming the era ("vintage travel poster") and the print texture does more than any adjective pile.
Infographic-style prompt examples
Clean infographic explaining the water cycle, flat vector
illustration style, labeled arrows showing evaporation and
rainfall, sans-serif labels with leader lines, sky blue and
white palette, generous spacing between elements
Why it works: "labels with leader lines" tells the model this is a diagram, not a painting with words in it.
Isometric infographic of a solar-powered house, cutaway showing
panels, battery, and wiring, numbered callouts 1-4 with short
sans-serif captions, soft 3D shading, white background, organized
grid layout
Why it works: numbered callouts plus captions keep text short — the regime where in-image text stays reliable.
Comparison infographic of three coffee brew methods, three columns
with simple line-art icons of each device, bold column headers
"POUR OVER", "FRENCH PRESS", "ESPRESSO", two-line descriptions
under each, clean editorial layout
Why it works: columns and headers define the layout before any text is drawn.
Bilingual signage prompt examples
This is GLM-Image's signature strength — its 91.16% score on the CVTG-2K Chinese text-rendering benchmark is exactly this job. Keep each language's copy short and state equal prominence when both must read as primary.
Café storefront sign, warm wooden board, "精品咖啡" in large
characters with "SPECIALTY COFFEE" in smaller letters directly
beneath, hand-lettered feel, soft morning light, photorealistic
Why it works: hierarchy is explicit — primary language, secondary line, stated sizes.
Bilingual conference banner, "科技峰会 2026" and "TECH SUMMIT 2026"
stacked with equal prominence, deep blue background, white modern
sans-serif typography, subtle circuit-board motif in the corners
Why it works: "equal prominence" prevents one language from shrinking into decoration.
Product packaging mockup for jasmine tea, vertical layout, elegant
"茉莉花茶" calligraphy as the hero element, "JASMINE TEA" in small
caps at the base, cream and jade green palette, minimalist design
Why it works: one hero text element per language, with roles assigned — calligraphy hero, small-caps support.
Airport wayfinding sign mockup, dark navy panel, white sans-serif
"出发 DEPARTURES" with English and Chinese side by side, small
arrow icon, flat photorealistic render, even studio lighting
Why it works: wayfinding is a low-word-count, high-clarity genre — exactly where bilingual rendering is dependable.
How to adapt these GLM-Image prompt examples without breaking them
Change one variable per generation. Swap the subject, keep the lighting; swap the lens, keep the palette. Each single change teaches you what that phrase actually controls, and any failure has exactly one suspect. Keep a note of your winners — a personal library of prompts that worked, with their outputs, beats any template list including this one. And when a result is close but wrong in one region, graduate to image-to-image: feed it back with "keep everything unchanged, change only X." The workflow is covered in the GLMImage walkthrough.
Frequently Asked Questions
Can I use these prompts with other image models?
Mostly yes — the subject/style/detail structure is universal. The text-bearing and bilingual examples are the exception: they lean on GLM-Image's text rendering, which is where it differs most from other generators.
Should I copy a prompt exactly as written?
Use it as a first run to see the pattern in action, then swap in your own subject and setting. The structure is the asset; the words are demonstrations.
Can these prompts include Chinese text in the image?
Yes — GLM-Image handles Chinese in-image text as a core strength and processes prompts in Chinese or English. Keep in-image copy to short phrases for the most reliable results.
How do I get more consistent results across a series?
Freeze everything except one variable: same prompt template, same style and lighting phrases, one subject swap per generation. Reuse the seed when you need near-identical baselines, then vary deliberately.
Start with the category closest to your work: open the generator → and paste.
