PromptMake
2026-06-25·15 min read

AI Art Style Prompts: From Impressionist to Brutalist — Cheat Sheet 2026

Copy-paste style tokens for Midjourney v7, FLUX, and DALL·E — Impressionism through Brutalism, hybrid combos, model picks, and how to avoid muddy style soup.

image-promptsmidjourneyfluxdall-eguidetips

Turn any photo into an AI prompt — free

No sign-up required. Works with Midjourney, FLUX, DALL-E.

Try Image to Prompt →

Style is the highest-leverage word in an image prompt. Change "photorealistic" to "impressionist painting" and the same subject becomes a different universe — brushwork, palette, edges, emotional register.

AI models learned style from captions. "Impressionist" compresses thousands of training examples into one token cluster: loose strokes, light over form, plein air mood. You don't need art history lecture notes — you need the right keywords and which model renders them best.

This cheat sheet covers major movements from Impressionism to Brutalism with copy-paste prompt fragments, hybrid rules, and 2026 model notes. Research-backed from current style glossaries (ImageToPrompt, PromptWritingStudio, Archez on architecture) and v7 prompting practice.

How to use this cheat sheet

Structure: [Subject + scene] + [style block from below] + [lighting if photo-adjacent] + [parameters]

Rules:

  1. Dominant style first — first style keyword carries most weight (Midjourney, FLUX, DALL·E agree on this).
  2. Max 2–3 style modifiers — more = muddy average.
  3. Assign roles when hybridizing — movement A for composition, B for color, C for texture.
  4. Go beyond the movement name — add technique + medium ("loose impasto brushwork" not just "impressionist").
  5. Save winners — build a personal Style Bible of strings that worked on your generator.

Style sits in element 5 of our 7-element visual framework; lighting (element 3) still matters — Baroque without chiaroscuro words underdelivers.

Model quick pick (same style, different renderer)

| Style type | Midjourney v7 | FLUX.1 | DALL·E 3 | SDXL + LoRA |

| Fine art historical | Excellent (richest vocab) | Good, less precise on niche | Very good | Checkpoint-dependent |

| Photoreal / realism | Good with --style raw | Excellent | Good | Excellent with photoreal checkpoints |

| Illustration / graphic | Excellent | Good | Excellent | Strong with anime/illustration LoRAs |

| Architecture (Brutalism etc.) | Strong | Excellent photoreal concrete | Good | Strong |

Midjourney often adds "house polish" — --style raw when you want literal style obedience over MJ gloss.

Fine art & traditional (1–10)

Impressionism

Visual: loose visible brushstrokes, soft edges, light and color over sharp form, plein air atmosphere.

Tokens: impressionist painting, loose impressionist brushwork, plein air, monet-style, visible dabs of pigment, dappled afternoon light

Snippet: Garden path in soft afternoon light, impressionist oil painting, loose brushstrokes, warm high-key palette, plein air atmosphere

Best: Midjourney, DALL·E 3

Post-Impressionism / Van Gogh adjacency

Visual: thicker swirling strokes, bolder color, emotional intensity.

Tokens: post-impressionist, thick impasto, swirling brushstrokes, expressive color, starry night palette

Snippet: Village street at night, post-impressionist oil, thick impasto swirling sky, bold yellow and blue

Expressionism

Visual: distorted forms, harsh angles, non-naturalistic color, psychological intensity.

Tokens: expressionist painting, german expressionism, distorted figures, agitated brushwork, munch-inspired mood

Snippet: Anxious figure on bridge, expressionist painting, distorted perspective, harsh red and black palette

Best: Midjourney, SD with expressionist checkpoint

Surrealism

Visual: dream logic, hyperreal impossible juxtapositions, precise render of absurd scenes.

Tokens: surrealist painting, dreamlike, magritte-style, dali-inspired, hyperrealistic surrealism

Snippet: Businessman with cloud head walking on ocean surface, surrealist painting, crisp detail, soft horizon

Best: Midjourney (native strength), DALL·E for conceptual

Cubism

Visual: fragmented geometry, multiple viewpoints, flat overlapping planes.

Tokens: cubist painting, analytical cubism, picasso-inspired, geometric fragmentation, muted cubist palette

Snippet: Portrait of violinist, analytical cubism, fractured planes, ochre and grey tones

Art Nouveau

Visual: flowing organic curves, botanical motifs, decorative borders, Mucha/Klimt elegance.

Tokens: art nouveau, mucha style, jugendstil, decorative floral borders, sinuous organic lines, klimt-inspired gold

Snippet: Woman with flowing hair surrounded by ivy ornaments, art nouveau poster, flat jewel tones, gold accents --ar 2:3

Best: Midjourney (exceptional)

Art Deco

Visual: 1920s geometry, symmetry, gold and black, sunburst and chevron motifs.

Tokens: art deco, 1920s art deco poster, geometric symmetry, gatsby-era, metallic gold accents

Snippet: Jazz age cocktail lounge interior, art deco design, symmetrical geometry, gold and emerald palette

Baroque

Visual: dramatic chiaroscuro, dynamic movement, rich warm palette, old master intensity.

Tokens: baroque painting, caravaggio lighting, rembrandt chiaroscuro, dramatic old master, tenebrism

Snippet: Still life with fruit and silver, baroque oil painting, single strong light source, deep shadows

See our cinematic lighting guide for chiaroscuro keywords.

Realism / Academic

Visual: unidealized truth, everyday subjects, natural light, 19th-century oil handling.

Tokens: 19th century realism, academic realism, naturalistic oil painting, unidealized portrait

Snippet: Farmer in field at dusk, realist oil painting, earth tones, natural overcast light

Watercolor

Visual: transparent washes, bleeding edges, paper texture, luminous whites.

Tokens: watercolor painting, loose watercolor illustration, wet-on-wet, visible paper texture, delicate wash

Snippet: Coastal cliff sketch, loose watercolor, soft bleeding blues, white paper showing through

Best: Midjourney, DALL·E, SD + watercolor LoRA

Modern & digital (11–20)

Oil painting (medium anchor)

Tokens: oil on canvas, thick oil paint, visible canvas texture, alla prima, glazing layers

Use when you want medium without a specific movement.

Pop Art

Visual: bold flat color, Ben-Day dots, commercial imagery, high contrast.

Tokens: pop art, lichtenstein-style dots, warhol-inspired, bold primary colors, comic panel aesthetic

Snippet: Celebrity portrait, pop art screen print, bold flat colors, halftone dots, high contrast

Minimalism

Visual: reduction, simple forms, limited palette, negative space.

Tokens: minimalist composition, reductive forms, sparse palette, large negative space

Snippet: Single red circle on off-white field, minimalist abstract painting, vast empty space

Ukiyo-e / Japanese woodblock

Visual: flat color planes, bold outlines, Edo-period composition.

Tokens: ukiyo-e woodblock print, hokusai-inspired, flat color waves, japanese edo period illustration

Snippet: Mountain and fishing boat, ukiyo-e woodblock print, flat indigo and cream palette

Concept art / matte painting

Visual: cinematic world-building, production design, atmospheric depth.

Tokens: digital concept art, matte painting, environment concept, cinematic key art

Snippet: Floating city above clouds, sci-fi concept art, atmospheric perspective, production design

Pixel art

Tokens: pixel art, 16-bit retro game, limited color palette, crisp pixel grid

Snippet: Medieval tavern interior, pixel art, warm torchlight, 32-color palette

Low poly / 3D stylized

Tokens: low poly 3D, stylized 3D render, faceted geometry, isometric game art

Vaporwave / Synthwave

Visual: 80s nostalgia, neon grids, pastel pink and cyan, retro futurism.

Tokens: vaporwave aesthetic, synthwave, neon grid sunset, retro 80s digital, chrome statue

Snippet: Palm trees against purple sunset grid, vaporwave aesthetic, neon pink and cyan

Cyberpunk

Visual: neon dystopia, rain-slick streets, holograms, high contrast.

Tokens: cyberpunk aesthetic, neon-lit alley, holographic ads, dystopian megacity, rain-slick pavement

Snippet: Hacker in neon alley, cyberpunk aesthetic, volumetric fog, magenta and cyan neon --ar 16:9

Hybrid tip: watercolor painting of cyberpunk city — dominant medium first (ImageToPrompt / PromptWritingStudio 2026 guidance).

Illustration & genre (21–30)

Studio Ghibli / anime soft

Tokens: studio ghibli style, anime background painting, soft cel shading, pastoral anime

Note: use style descriptors; avoid living artist names when policies block.

Manga / ink

Tokens: black and white manga, ink line art, screentone shading, dynamic action lines

Comic book (American)

Tokens: american comic book, bold ink outlines, cel shading, superhero comic panel

Children's book illustration

Tokens: children's book illustration, whimsical storybook, soft rounded shapes, gentle pastel palette

Botanic / scientific illustration

Tokens: botanical illustration, scientific plate, precise linework, watercolor specimen study

Art Brut / naive art

Tokens: naive art, folk art painting, childlike perspective, primitive flat composition

Graffiti / street art

Tokens: graffiti mural, spray paint street art, bold tags, urban wall texture

Steampunk

Tokens: steampunk aesthetic, brass and copper machinery, victorian industrial, clockwork gears

Hybrid: steampunk + art nouveau — shared ornamental logic (AI Arsenal style-control notes).

Solarpunk

Tokens: solarpunk aesthetic, green futuristic architecture, optimistic eco city, warm natural light

Dark fantasy illustration

Tokens: dark fantasy art, grim illustration, muted palette, ornate gothic details

Photography-adjacent styles (31–38)

Use --style raw on Midjourney for these when photographic.

Film photography

Tokens: shot on 35mm film, kodak portra 400, cinestill 800t, film grain, analog color

Documentary

Tokens: documentary photography, candid moment, natural available light, photojournalism

Fashion editorial

Tokens: high fashion editorial, vogue-style portrait, studio fashion photography, bold styling

Macro / product

Tokens: macro product photography, studio softbox, white seamless, catalog shot

Infrared / experimental photo

Tokens: infrared photography, false color foliage, surreal landscape photo

Tilt-shift miniature

Tokens: tilt-shift photography, miniature effect, selective blur, diorama look

Double exposure

Tokens: double exposure photography, silhouette blended with forest, layered portrait

Long exposure

Tokens: long exposure, light trails, smooth water mist, night city streaks

Architecture & design (39–46) — including Brutalism

Brutalism

Visual: raw exposed concrete (béton brut), monolithic massing, heavy horizontal lines, deep shadow, cantilevered volumes.

Tokens: brutalist architecture, raw board-formed concrete, monolithic brutalist form, cantilevered concrete volumes, soviet brutalist monument

Snippet: Civic library on overcast day, brutalist architecture, exposed concrete texture, deep shadow lines, minimal windows --ar 3:2 --style raw --v 7

Archez International (2026): add exposed board-formed concrete and cantilevered volumes to prevent drift to generic modernism.

Best: Midjourney, FLUX for photoreal concrete

Bauhaus / Modernism

Tokens: bauhaus design, modernist architecture, functional geometry, primary color accents

Parametric / Zaha-adjacent

Tokens: parametric architecture, fluid curvilinear building, continuous white surfaces

Note: training bias toward ZHA — specify alternate cues if you want non-Zaha curves.

Scandinavian minimal interior

Tokens: scandinavian interior design, light oak, hygge, minimal furniture, soft daylight

Industrial / loft

Tokens: industrial interior, exposed brick, steel beams, warehouse conversion

Art Deco architecture

Tokens: art deco building facade, stepped ziggurat tower, 1920s geometric ornament

Japanese contemporary (Tadao-adjacent)

Tokens: minimal japanese architecture, concrete and light, zen courtyard, clean lines

Biophilic / organic architecture

Tokens: biophilic design, living green facade, organic architecture, nature-integrated building

Brutalism ↔ Impressionism (intentional hybrid example)

Combining opposing styles works when roles are split (GetVidzy 2026):

Brutalist concrete tower rising from impressionist landscape — building rendered with raw concrete and hard shadow, surrounding hills in loose monet-style brushwork, oil painting hybrid

Don't: "impressionist brutalist" alone — model averages into grey mush.

Style weighting by platform

Midjourney: art nouveau::2 cyberpunk::1 in prompt; --stylize 250–400 for painterly styles, lower with --style raw for photo styles.

SDXL: (impressionism:1.2) sparingly; negative field blocks wrong medium.

FLUX: dominant style in first clause; no :: syntax.

DALL·E: "in the style of [movement]" mid-paragraph; revise in chat.

What to avoid

Listing 5+ movements ("impressionist cubist art deco cyberpunk watercolor")

Style without subject — model generates generic demo scene

Living artist names when platform blocks — use movement + technique instead

Same string for MJ and FLUX without translation

Ignoring medium conflict: "photorealistic watercolor oil sketch"

PromptMake workflow

/image → target Midjourney v7 or FLUX → Change Style mode: keep composition, swap movement tokens.

/image → Recreate from reference art to extract style vocabulary.

3 free generations/day on /image. Build Style Bible from outputs.

Frequently asked questions

Best style for beginners?

Impressionism, Art Nouveau, Pop Art — distinctive, forgiving (ImageToPrompt glossary).

Brutalism for characters?

Brutalism is architectural — pair with separate character illustration style or place figure in brutalist setting.

Do artist names still work?

Sometimes; movement + technique is safer and more controllable in 2026 policies.

How many styles in one prompt?

Two max for hybrids; three only with assigned roles (composition/color/texture).

Style vs lighting order?

Style medium early; lighting phrases after subject (see cinematic lighting guide).

Related articles

7 elements of visual prompts — where style fits.

Cinematic lighting keywords — Baroque, noir, golden hour.

Midjourney prompts 2026 — parameters for style control.

DALL·E vs MJ vs Flux — which model needs better style dialect.

Reverse-engineer prompt from photo — steal style from references.

Bottom line

Art movements are compressed prompt libraries. This cheat sheet runs Impressionism → Brutalism with tokens you can paste, model picks, and hybrid rules.

Pick one dominant style, name the technique not just the movement, keep subject specific, translate per platform. Save what works — your Style Bible beats any generic list after ten sessions.

Ready to generate your own prompts?

Free. No sign-up required. Works with all major AI models.

Related articles