How to Write Prompts for Grok Imagine: Complete Guide with 20 Copy-Paste Examples
Grok Imagine is xAI's image generation feature, available inside Grok on X (formerly Twitter) and the Grok web app. Under the hood it runs FLUX.1 — Black Forest Labs' architecture — which is one of the most capable open-weight image models available. But FLUX.1 responds to prompts differently than MidJourney, DALL-E, or Stable Diffusion, and the guides written for those tools won't get you the best results here.
This guide covers the exact prompt anatomy that works with Grok Imagine, what to avoid, and 20 copy-paste prompts you can use right now.
---
How Grok Imagine Works
Grok Imagine uses FLUX.1 (specifically a fine-tuned variant). FLUX.1 was trained on a broad dataset with strong emphasis on:
It does not use MidJourney-style suffix parameters like --ar 16:9 or --style raw. You describe everything in natural language.
---
Prompt Anatomy: The FLUX.1 Formula
The highest-performing prompts for Grok Imagine follow this structure:
[Subject] + [Action/pose] + [Setting/environment] + [Lighting] + [Camera/lens] + [Style/mood] + [Quality tag]
Example: > "A young woman reading a book in a sunlit café, seated by a rain-streaked window, warm golden backlight creating a halo effect, shallow depth of field, shot on 85mm f/1.4, cinematic mood, film grain, highly detailed"
Each element adds a layer. Skipping the lighting description is the single most common mistake — FLUX.1 responds very specifically to lighting cues.
---
What Works (and What Doesn't)
Works
Doesn't Work
--ar, --chaos, --style)---
20 Copy-Paste Prompts for Grok Imagine
Portraits
1. Cinematic portrait > "A close-up portrait of a middle-aged man with weathered skin and kind eyes, dramatic side lighting, shallow depth of field, film grain, desaturated earth tones, highly detailed, 50mm lens"
2. Fantasy character > "A female elf warrior in silver armor standing in an enchanted forest at dusk, bioluminescent flowers around her feet, cool blue and warm amber lighting contrast, fantasy epic style, highly detailed"
3. Studio-quality headshot > "Professional headshot of a woman in her 30s with natural makeup, white studio background, softbox lighting, clean and sharp, editorial photography style"
Environments
4. Urban night scene > "Aerial view of a neon-lit Tokyo street crossing at 2 AM, rain-soaked pavement reflecting pink and blue signage, long exposure light trails, cinematic widescreen, 24mm lens"
5. Forest atmosphere > "Ancient redwood forest at golden hour, shafts of light breaking through the canopy, ground fog drifting between the roots, wide angle, high dynamic range, National Geographic style"
6. Interior design > "A minimalist Japandi living room at dawn, warm morning light streaming through floor-to-ceiling windows, neutral linen textures, bonsai on a low wooden table, architectural photography"
Products
7. Product photography > "A luxury perfume bottle on a black marble surface, dramatic spotlight from above, sharp reflections, water droplets, dark and elegant, commercial photography style"
8. Food photography > "A bowl of ramen in a Tokyo restaurant, steam rising, perfectly arranged toppings, warm overhead light, tight macro shot, food magazine editorial style"
Artistic Styles
9. Impressionist landscape > "A lavender field in Provence at sunset, painted in the style of Monet, loose impressionist brushstrokes, warm gold and purple palette, dreamy and atmospheric"
10. Retro poster > "1960s travel poster style illustration of Rome, Colosseum in background, warm oranges and teals, bold serif typography reading 'ROMA', flat graphic design"
11. Cyberpunk scene > "A cyberpunk street market in 2087, holographic vendor stalls, rain, steam rising from grates, a lone figure in a neon-lit rain jacket walking away from camera, ultra detailed, atmospheric"
12. Vintage photography > "A 35mm film photograph from the 1970s of a woman at a roadside diner, faded color, grain, slight vignette, Kodachrome color profile, nostalgic"
Characters
13. Chibi character > "A chibi-style anime character of a small wizard holding an oversized spellbook, big expressive eyes, pastel colors, clean line art, white background, cute and detailed"
14. Action figure > "A collectible action figure of a sci-fi space explorer in a clear plastic clamshell box, product photography on white background, manufacturer branding on packaging, hyper realistic"
15. Pixar character > "A Pixar-style animated character of a cheerful baker in a colorful apron, round and expressive face, warm studio lighting, CGI render quality, family-friendly"
Nature
16. Ocean macro > "A macro photograph of a sea anemone in a tidal pool, vivid orange and purple colors, crystal clear water, sunlight refracting through shallow water, sharp focus, National Geographic quality"
17. Wildlife > "A snow leopard resting on a rocky mountain ledge at sunrise, golden light on its spotted fur, mist in the valley below, telephoto lens compression, wildlife documentary photography"
Text-in-Image
18. Logo mockup > "A coffee shop logo on a kraft paper cup, hand-drawn vintage style illustration of a mountain with a coffee bean, warm brown and cream tones, the text reads 'Summit Coffee Co.' in bold slab serif"
19. Poster with text > "A minimalist movie poster, dark background, a silhouette of a person standing in a doorway with bright light beyond, the text 'THE WAY OUT' centered in large bold white Helvetica at the top"
20. Social media graphic > "A clean motivational quote card, sage green background, white serif text reading 'Start before you're ready.', small botanical illustration of a fern in the corner, modern editorial design"
---
Grok Imagine vs Other Tools: When to Use Which
| Use Case | Best Tool | |---|---| | Photorealistic portraits | Grok Imagine (FLUX.1), GPT Image 2.5 | | Text-in-image | Grok Imagine, Ideogram 4.5 | | Fantasy / stylized art | MidJourney V8 | | Product photography | Reve 2.0, Grok Imagine | | Chibi / anime style | MidJourney V8, Grok Imagine | | Logo / graphic design | Ideogram 4.5, Grok Imagine |
Grok Imagine's FLUX.1 base makes it exceptional for photorealism and text. For heavily stylized or artistic outputs, MidJourney V8 still leads.
---
Advanced Tips
Tip 1: Use negative space in descriptions
Instead of "a person NOT wearing glasses" (negations are unreliable), describe what you DO want: "a person with clear eyes, no accessories."Tip 2: Stack quality tags at the end
End every prompt with 2–3 quality anchors: "highly detailed," "professional photography," "8K," "award-winning." These shift the model toward its highest-quality output range.Tip 3: Specify the format
FLUX.1 defaults to square images. If you want portrait or landscape: "vertical 9:16 format" or "wide 16:9 cinematic aspect" works in natural language.Tip 4: Iterate with small changes
Grok Imagine lets you regenerate. Change one element at a time (swap "golden hour" for "blue hour," keep everything else) to understand what each element contributes.Tip 5: Use the chibi / toyification shortcut
"Chibi style" and "action figure style" are strong activation patterns in FLUX.1. Combining one of these with a character description almost always produces clean, on-brand results.---