Google Gemini Image Prompting Guide
Google Gemini Image is the image-generation side of the Gemini family — three tiers (Standard at 1 credit, Pro at 4.5-7.5 credits depending on resolution), 1K/2K/4K output, 10 aspect ratios including 21:9 cinematic and 4:5 social-spec.
Google Gemini Image is the image-generation side of the Gemini family. Three model tiers from one tool: Standard (1 credit, Gemini Flash Image) for everyday work, Pro (4.5-7.5 credits depending on resolution) for hero shots, and a second Pro variant (Nano Banana 2) at lower pricing: 2.25 credits at 1K, 3 at 2K, 4.5 at 4K.
Three resolution tiers (1K, 2K, 4K) — only Gemini Image among general models supports native 4K. Ten aspect ratios cover everything from cinematic 21:9 through social-spec 4:5 (Instagram portrait) to ultra-vertical 9:16. Standard tier is plenty for iteration; Pro for material-driven hero shots.
3
Model tiers
10
Aspect ratios
4K
Native max resolution
Gemini Flash Image — the Standard tier — runs at 1 credit per image and produces strong general-purpose output. It's the cheapest tier across PixelDojo's premium image models and surprisingly competitive on most subjects. Default starting point.
Pro tier trades 1 credit for 4.5 (1K, 2K) or 7.5 (4K) in exchange for stronger fidelity — better skin texture, sharper material edges, more refined lighting. Use Pro for hero shots and material-driven scenes; Standard is plenty for iteration.
Pro tier supports native 4K — true 16MP output, not upscaled 2K. Most general image models max out at 2K or upscale internally. Useful for print-spec, very-large-format display, and any output where 4K compositing fidelity matters.
Gemini Image commits to non-photo aesthetics (illustration, watercolor, pastel) cleanly when prompted. Same model that handles photoreal landscapes also produces strong storybook and editorial illustration without needing aesthetic-specific tooling.
Each example shows the exact prompt that produced the result. Copy any prompt with one click.
Standard · 2K · 21:9 · 1 credit
Cinematic establishing shot of a misty Vietnamese coastal village at dawn, traditional fishing boats moored at the dock, lanterns still glowing in the soft pre-sunrise light, painterly photoreal, atmospheric
21:9 ultra-wide + cultural specificity + atmospheric anchor = strong cinematic landscape from Standard tier at 1 credit. Gemini Image composes 21:9 well — pictorial depth, strong horizon, atmospheric haze all read cleanly without needing Pro tier fidelity.
Standard · 2K · 3:4 · 1 credit
Editorial portrait of an elderly artisan ceramicist with weathered hands holding a freshly-thrown vase, soft warm afternoon studio light, clay dust on his apron, photoreal documentary style
"Documentary style" anchors the framing toward naturalistic composition over fashion-editorial setups. Weathered-hands / clay-dust specificity gives Gemini concrete texture targets. Standard tier holds up on documentary portraits — reach for Pro on fashion editorial.
Standard · 2K · 4:5 · 1 credit
Hyper-real macro of fresh pomegranate seeds spilling from a halved fruit onto a dark slate surface, droplets of juice catching cool side lighting, food editorial composition
4:5 is Instagram portrait-spec — the social-native ratio for vertical-feed posts. Pair with food-editorial composition language and Gemini Image renders fruit pulp, juice droplets, and slate surface texture cleanly. Crop-free output for the social channel.
Standard · 2K · 1:1 · 1 credit
Stylized digital illustration of a koi pond at twilight viewed from above, lily pads with bioluminescent flowers, two orange koi circling, soft pastel palette, dreamy storybook quality
Stylized work needs explicit medium ("digital illustration", "soft pastel palette", "storybook quality") to commit Gemini out of photoreal default. Top-down compositions ("viewed from above") work well — Gemini composes the layout convincingly.
Lite · 1K · 2:3 · 2 credits
A vintage-style travel poster for Kyoto, Japan. Bold clean headline typography reading 'KYOTO' across the top and 'LAND OF TRADITION' beneath it, a red pagoda among cherry blossoms with Mount Hiei in the background, flat screen-print illustration, limited warm palette, crisp legible lettering.
Nano Banana 2 Lite is the cheapest path to the Nano Banana family’s standout text rendering — pick it from the model dropdown for 1K work at 2 credits. Put the exact words in quotes and name the layout ("headline across the top"); legible typography is where this family shines.
Gemini Flash Image (Standard, 1 credit) is the cheapest tier across PixelDojo's premium image models and surprisingly strong. Use it for iteration and most production work. Move to Pro only when you need the marginal extra fidelity on a hero shot.
The model dropdown also has Nano Banana 2 Lite — a faster, lower-cost member of the Nano Banana family at 2 credits, fixed at 1K. It keeps the family's strong typography and prompt-following, so it's a great pick for quick iterations, social-size graphics, and text-forward posters where you don't need 2K/4K.
4K (Pro, 7.5 credits) is true 16MP — useful for print-spec or large-format display. For social, web, and most digital use, 2K is more than enough. Don't burn 7.5 credits on output that'll be downscaled to 1080px wide.
4:5 is the Instagram portrait spec (1080×1350) — the social-native ratio for vertical-feed posts that take up more screen space than 1:1. Generate at 4:5 for crop-free social output. Same recipe works for 5:4 horizontal Facebook posts.
Gemini Image handles cultural specificity ("Vietnamese coastal village", "Mongolian eagle hunter", "Tibetan monastery") with strong atmosphere and respectful realism. Documentary-style portraits also read well — use "documentary style" as an anchor.
For stylized work, name the medium ("digital illustration", "watercolor", "pastel"), the tradition ("storybook", "editorial illustration"), and the palette. Gemini commits cleanly when the prompt does too — vague "stylized" produces photo-illustration hybrids.
21:9 reads cinematic — pictorial, horizon-driven. 4:5 reads social-native — Instagram portrait. 1:1 for square. 9:16 for vertical reels. Generate at the right aspect to match downstream channel rather than cropping later.
| Setting | Values | Notes |
|---|---|---|
| Model tier | Standard · Pro · Pro (FAL v2) | Standard at 1 credit. Pro at 4.5-7.5 credits by resolution. |
| Resolution | 1K · 2K · 4K | Only Gemini Image supports native 4K among general models. |
| Aspect ratio | 1:1 · 2:3 · 3:2 · 3:4 · 4:3 · 4:5 · 5:4 · 9:16 · 16:9 · 21:9 | Ten presets including social-spec 4:5 and cinematic 21:9. |
| Output format | PNG · JPEG · WEBP | PNG default. Pick by downstream use. |
| Pricing (Standard) | 1 credit any resolution | Cheapest premium image tier on PixelDojo. |
| Pricing (Pro) | 4.5 cr (1K, 2K) · 7.5 cr (4K) | Resolution-scaled. Reserve for hero shots. |
They're related — Nano Banana Pro is a separate tool that uses one of the Gemini Image Pro variants under the hood, with a focused workflow around the 'banana' aesthetic. Google Gemini Image gives you the full multi-tier surface (Standard, Pro variants) for general work; Nano Banana Pro is the streamlined product.
Standard (1 credit, Gemini Flash Image) is the default — cheap, fast, strong on most subjects. Use Pro (4.5-7.5 credits) for hero shots, material-driven scenes (food, skin, fabric texture), and any time you want the marginal extra fidelity. Standard handles 80%+ of production work.
Two Gemini Image Pro backend variants with slightly different pricing curves — one runs 4.5 credits at 1K and 2K, 7.5 credits at 4K; the other runs 2.25 credits at 1K, 3 at 2K and 4.5 at 4K (cheaper at every resolution). PixelDojo exposes both so you can pick by cost profile.
Native 16-megapixel output — true 4K-spec dimensions, not upscaled 2K. Most general image models cap at 2K or upscale internally; Gemini Pro is one of the few with native 4K. Useful for print-spec, very-large-format display, or any downstream pipeline where 4K compositing matters.
Gemini renders short text reasonably well — signs, labels, packaging. Pro tier holds typography better than Standard. For text-heavy compositions (multi-line headlines, paragraphs), prefer FLUX Kontext Pro, Recraft Pro, or Qwen Image. Quote exact strings in double quotes.
Yes. Multi-tier exposure: Gemini Image (Standard) appears in the Canvas model picker as the default Gemini option, with Pro variants accessible too. Public API endpoint is /api/v1/models/google-gemini-image/run with prompt, modelTier, resolution, aspect_ratio, and output_format fields.