Ahegao face prompts for Nano Banana work best when you specify the expression mechanics (rolled eyes, tongue position, blush intensity) alongside camera, lighting, and aspect ratio. The model responds to concrete facial geometry, not vague mood words. Below are copy-paste prompts and the exact levers that keep results consistent across nano banana 2 aheago face, nano banana pro aheago face, and standard nano banana aheago face runs.
Why vague prompts fail for this expression
Nano Banana-class models (Gemini image generation) treat "ahegao" as a high-level tag. If you stop there, you get a generic exaggerated face: eyes rolled, mouth open, but no control over pupil position, tongue shape, sweat placement, or skin flush. The fix is to describe the face as a set of coordinates and states. Think of it as directing a portrait photographer, not writing a mood board.
Three variables drive 80% of the output:
- Eye state — pupil position (rolled up, crossed, half-lidded), eyelid height, and whether the whites are visible above or below the iris.
- Mouth and tongue — open width, tongue extension (curled, flat, pointed), saliva or gloss highlights.
- Skin response — blush spread (cheeks only vs. nose bridge vs. ears), sweat beads, and micro-tremor lines.
Add camera and lighting, and the prompt becomes repeatable. For a broader expression-control framework, see portrait expressions and emotions control.
Core prompt structure
Use this order: subject → expression mechanics → skin response → camera → lighting → aspect ratio → style anchor.
Example:
anime girl, close-up face, eyes rolled upward with pupils hidden under upper eyelids, mouth open wide, tongue extended and curled at the tip, deep blush across cheeks and nose bridge, three sweat beads on forehead, 85mm portrait lens, soft ring light from above, 2:3 aspect ratio, cel-shaded anime style
That prompt works on nano banana aheago face and nano banana 2 aheago face without changes. For nano banana pro aheago face, you can add a second clause for detail density (pores, eyelash separation, fabric texture) because the pro tier resolves finer grain.
Copy-paste prompt examples
1. Classic anime close-up (2:3)
anime girl, close-up face, eyes rolled up and crossed, pupils barely visible, eyelids half-closed, mouth open in a wide O, tongue extended flat, heavy blush on cheeks and ears, two sweat drops on temple, soft diffused studio light, 85mm lens, shallow depth of field, cel-shaded anime style, 2:3 aspect ratio
2. Manga ink style (1:1)
manga panel, girl face, eyes rolled back with visible whites, mouth open, tongue out, blush rendered as cross-hatch lines, sweat drops as solid black shapes, high-contrast ink, screentone shading, 1:1 aspect ratio
3. Semi-realistic portrait (3:4)
semi-realistic anime portrait, young woman, eyes rolled upward, pupils hidden, lower eyelids raised, mouth open, tongue curled, glossy saliva highlight on lower lip, blush gradient from cheek to ear, skin texture with visible pores, 50mm lens, soft window light from left, 3:4 aspect ratio
4. Neon-lit night scene (16:9)
anime character, face close-up, ahegao expression, eyes rolled up, tongue out, blush under neon pink and cyan rim light, sweat beads catching light, dark background with bokeh, 35mm lens, 16:9 aspect ratio, cyberpunk color palette
For more neon portrait setups, see neon light portrait prompts.
5. Chibi / stylized (1:1)
chibi anime girl, oversized head, ahegao face, eyes as spiral swirls, mouth wide open, tongue out, blush as two pink circles, sweat drop on forehead, flat pastel background, 1:1 aspect ratio, vector-like clean lines
6. Dynamic angle (2:3)
anime girl, low-angle close-up, ahegao expression, eyes rolled up, mouth open, tongue extended, blush and sweat, hair blown back, motion lines, 24mm lens, dramatic underlighting, 2:3 aspect ratio
Tuning for nano banana 2 and pro
Nano Banana 2 and Pro differ in how much fine detail they retain. On nano banana 2 aheago face, keep the prompt tight: one expression clause, one lighting clause, one camera clause. On nano banana pro aheago face, you can stack detail: eyelash separation, iris reflection, saliva strands, fabric weave. If the pro output looks over-sharpened, drop the lens mm to 50 and add "soft focus on skin."
For a full breakdown of model differences, see Nano Banana 2 prompts and Nano Banana Pro use cases.
Lighting and camera levers
Lighting changes the read of the expression. Hard light makes sweat and saliva pop; soft light makes blush blend. Camera distance controls how much face fills the frame. A 85mm lens gives compression and flattering proportions; 24mm gives distortion and energy.
- Blush visibility: soft frontal light + 85mm.
- Sweat highlights: hard side light + 50mm.
- Tongue gloss: small specular light from below.
- Eye roll clarity: light from above so upper eyelids cast shadow over pupils.
For a deeper lighting guide, see mastering AI portrait lighting prompts and portrait lighting techniques.
Skin, sweat, and blush control
Blush is not one value. Specify spread: cheeks only, cheeks + nose bridge, cheeks + ears, or full face. Sweat can be beads, streaks, or sheen. Saliva can be a single strand, a gloss on the lip, or a drop on the chin. Each choice changes the tone from comedic to intense.
Example clause:
blush concentrated on cheeks and nose bridge, three distinct sweat beads on forehead, glossy saliva strand from lower lip to chin
For skin texture details, see portrait skin texture and makeup.
Aspect ratio and composition
- 1:1 — social posts, manga panels, chibi.
- 2:3 — portrait orientation, classic anime close-up.
- 3:4 — semi-realistic, print-friendly.
- 16:9 — cinematic, neon night scenes.
Keep the face centered or rule-of-thirds. For composition rules, see AI image composition techniques.
Negative prompts and safety
Use negative prompts to remove artifacts: "no extra fingers, no distorted eyes, no blurry tongue, no asymmetric pupils." For a full list, see negative prompts guide.
Also, keep outputs within platform guidelines. Ahegao is an exaggerated expression; avoid prompts that imply explicit content. Focus on face mechanics, not body. If a model refuses, simplify the prompt and remove ambiguous terms.
Style anchors
Anchor the style to keep consistency across runs: "cel-shaded anime," "manga ink," "semi-realistic anime," "chibi vector." For more anime styles, see anime and manga style. For broader style fusion, see AI style fusion techniques.
Workflow tips
- Start with a base prompt from above.
- Change one variable at a time (eye state, then mouth, then blush).
- Save the prompt that works as a template.
- For nano banana pro aheago face, add detail clauses last.
- Use the same seed if the model supports it to compare changes.
For a structured prompt syntax, see prompt structure and syntax guide.
Common pitfalls
- Too many adjectives: "beautiful, stunning, amazing" adds nothing. Use concrete nouns.
- Missing camera: without lens and distance, the model guesses.
- Ignoring aspect ratio: 16:9 will crop the face differently than 2:3.
- Overloading pro: more detail clauses can cause artifacts if the base is weak.
Example: full prompt for nano banana pro aheago face
semi-realistic anime portrait, young woman, close-up face, eyes rolled upward with pupils hidden under upper eyelids, lower eyelids raised, mouth open wide, tongue extended and curled at tip, glossy saliva strand from lower lip, blush gradient from cheeks to ears, three sweat beads on forehead, visible skin pores, individual eyelashes, 85mm lens, soft ring light from above, 3:4 aspect ratio, cel-shaded anime style with subtle texture
That prompt uses pro-tier detail without breaking the expression.
Final note
Ahegao prompts are a stress test for expression control. If you can direct eye roll, tongue position, blush spread, and sweat placement, you can direct any facial expression. Use the examples above as templates, then adjust one variable at a time.
Ready to skip the trial and error? Browse the ready-to-use prompt library at NanoPrompts for copy-paste templates across portraits, headshots, style transfers, and 3D collectible figures.
