Conclusion
bunny girlis the most stable.playboy bunnyhas high variance in outfit form.- Material specification is clearly reflected. The differences between patent leather (mirror gloss), velour (irregular sheen), and latex (wet gloss) are especially pronounced.
- Color specification is highly accurate. Red, pink, white, and gold are all reproduced precisely.
- Two-tone colors are also reproducible. Left-right color split comes out accurately.
- Bunny tail appearance rate is low. Even when explicitly specified, roughly 1 out of 3 images.
- Tailcoat jacket is wearable with
leotard under open tuxedo jacket. The key is specifying the layering relationship concretely (“leotard underneath, jacket on top”).
What This Article Covers
- The default style that comes out with “playboy bunny” and “bunny girl” base keywords
- How outfit changes with material specification (satin, patent leather, velour, latex, lace)
- How color specification affects bunny girl outfits
- Appearance rate of individual parts (bunny ears, bow tie, cuffs, tail)
Experiment Conditions
| Item | Value |
|---|---|
| Model | z-image-turbo (6B, photorealism-focused distilled model) |
| Steps | 8 |
| Sampler | euler |
| Scheduler | ddim_uniform |
| CFG | 1.0 |
| Image size | 1024×1024 |
| Seed | Fixed to 42, 123, 789 |
Base Prompt
Experiment 1: Base Keyword Comparison
Condition A: playboy bunny (2 tokens)
| seed=42 | seed=123 | seed=789 |
|---|---|---|
![]() | ![]() | ![]() |
Observations
The outfit form varies significantly across 3 images. seed42: white satin leotard with frills, white ears, black bow tie; seed123: black satin leotard, black ears; seed789: black bikini type (bra + thong rather than leotard) with black ears and bow tie. No tail visible in any of the 3. playboy bunny alone has low outfit consistency.
Condition B: bunny girl (2 tokens)
| seed=42 | seed=123 | seed=789 |
|---|---|---|
![]() | ![]() | ![]() |
Observations
All 3 images are stably in a black leotard form. Ears, bow tie, and cuffs are present in all 3. The leotard material has a satin or patent sheen. Ear color varies slightly (white ears 2/3, black ears 1/3). No tail confirmed in any. bunny girl is more stable than playboy bunny.
Condition C: bunny girl costume (3 tokens)
| seed=42 | seed=123 | seed=789 |
|---|---|---|
![]() | ![]() | ![]() |
Observations
All 3 have black leotard with ears, bow tie, and cuffs — stability equivalent to Condition B. Material is consistently satin across all 3. Stocking type varies (fishnet / white stockings), but the leotard body shape is consistent. Adding “costume” makes no practical difference.
Experiment 2: Classic Style — Part-by-Part Composition
Condition D: Classic Full Set Specification
| seed=42 | seed=123 | seed=789 |
|---|---|---|
![]() | ![]() | ![]() |
Observations
All 3 have black satin leotard with white cuffs, bow tie, and pantyhose — the highest accessory reproduction rate of all conditions. The high-cut leotard shape matches across all 3. Nude-tone pantyhose is accurately reproduced. The tail is visible as a white round puff in seed789 (1/3), but even with many parts specified, stable tail output remains difficult.
Experiment 3: Material Variations
Condition E: Satin
| seed=42 | seed=123 | seed=789 |
|---|---|---|
![]() | ![]() | ![]() |
Observations
All 3 in black satin with a soft, non-mirror sheen. Ears, bow tie, cuffs, and pantyhose are present in all 3. seed42 alone has a white/light-colored leotard (1/3). Satin’s characteristic soft curved highlights, no specular reflection — the most gentle sheen of all material conditions.
Condition F: Patent Leather (Enamel)
| seed=42 | seed=123 | seed=789 |
|---|---|---|
![]() | ![]() | ![]() |
Observations
All 3 in black patent leather / enamel with strong mirror gloss and sharp highlights visible on the surface. Ears are also the same black enamel material. The difference from satin (Condition E): the reflections on the leotard surface are sharper with whiter highlights — clearly distinct.
Condition G: Velour
| seed=42 | seed=123 | seed=789 |
|---|---|---|
![]() | ![]() | ![]() |
Observations
All 3 clearly render the velour/velvet texture, with the irregular sheen pattern characteristic of crushed velour visible on the surface. Color tends toward navy or grey-purple. Bow tie, ears, and cuffs are also consistently velour. The highest material specification fidelity of all conditions.
Condition H: Latex
| seed=42 | seed=123 | seed=789 |
|---|---|---|
![]() | ![]() | ![]() |
Observations
All 3 in black latex with an even stronger sheen than patent leather (Condition F), plus a body-hugging fit suggesting skin contact. The surface is a wet-looking mirror finish, with high body line conformity. F and H are distinguishable by gloss type — H has a wider reflection range with smoother highlight edges.
Condition I: Lace
| seed=42 | seed=123 | seed=789 |
|---|---|---|
![]() | ![]() | ![]() |
Observations
All 3 in lace leotard with floral lace transparency clearly expressed. Unlike other material conditions, color variance is higher (white 2/3, black 1/3). Ears tend toward fur material, and the lace material specification shifts the entire tone of the bunny girl look.
Experiment 4: Color Variations
Condition J: Red Bunny
| seed=42 | seed=123 | seed=789 |
|---|---|---|
![]() | ![]() | ![]() |
Observations
All 3 in red satin leotard with high color accuracy. Ears and bow tie are also red, cuffs are white. seed42 alone generated two people (artifact). Red saturation is vivid and satin sheen is consistent.
Condition K: Pink Bunny
| seed=42 | seed=123 | seed=789 |
|---|---|---|
![]() | ![]() | ![]() |
Observations
All 3 in pink satin leotard, consistently reproducing a light baby pink. Ears and bow tie are also pink. Overall pastel tone is unified and color consistency is high.
Condition L: White Bunny
| seed=42 | seed=123 | seed=789 |
|---|---|---|
![]() | ![]() | ![]() |
Observations
All 3 in white satin leotard, with ears showing white with visible pink inner satin/fur. seed789 alone changes to a cami type with straps (1/3). Pantyhose is white or skin-tone — not black pantyhose, which differs from other color conditions. When going fully white, the pantyhose also gets pulled toward white.
Condition M: Gold Bunny
| seed=42 | seed=123 | seed=789 |
|---|---|---|
![]() | ![]() | ![]() |
Observations
All 3 in gold metallic sheen leotard — very high color accuracy. Ears and bow tie also gold. The leotard surface has a strong metallic/glitter reflection unlike any other material condition. Pantyhose is black or skin-tone.
Condition N: Two-Tone Color
| seed=42 | seed=123 | seed=789 |
|---|---|---|
![]() | ![]() | ![]() |
Observations
All 3 have the leotard split into black and pink on left and right — the specification is accurately reflected. The dividing line is approximately centered vertically. Ears are also two-tone (black + pink). Bow tie and cuffs are present in all 3. Sufficient reproduction accuracy for a left-right two-tone design.
Experiment 5: Optional Parts
Condition T: Bunny Suit + Tailcoat (Name Specification)
| seed=42 | seed=123 | seed=789 |
|---|---|---|
![]() | ![]() | ![]() |
Observations
All 3 are set in a bar counter background with the basic set of black satin leotard, ears, bow tie, cuffs, and black pantyhose. However, the tailcoat jacket is not worn in any of the 3 — the prompt specification was not reflected. “tailcoat” may have influenced the background atmosphere (formal bar) instead.
Condition T8: Bunny Suit + Tailcoat (Layering Description)
Instead of wearing tailcoat jacket over leotard, an approach that writes the leotard first and then concretely describes the wearing state as under open ... tuxedo jacket.
| seed=42 | seed=123 | seed=789 |
|---|---|---|
![]() | ![]() | ![]() |
Observations
3 out of 3 successfully wore the tuxedo jacket. seed42: fully worn and closed, seed123: one shoulder slightly off in a sexy pose, seed789: also with sleeves through. The “bunny girl + tailcoat” style was stably reproduced. The key: using leotard under open ... tuxedo jacket to explicitly state the layering relationship with “leotard underneath, jacket on top.”
Cross-Comparison
Base Keyword Comparison (playboy bunny / bunny girl / bunny girl costume)
| Seed | playboy bunny (A) | bunny girl (B) | bunny girl costume (C) |
|---|---|---|---|
| 42 | ![]() | ![]() | ![]() |
| 123 | ![]() | ![]() | ![]() |
| 789 | ![]() | ![]() | ![]() |
bunny girl (B) is the most stable. playboy bunny (A) has high outfit variance.
Material Comparison (Satin / Patent Leather / Velour / Latex / Lace)
| Seed | Satin (E) | Patent Leather (F) | Velour (G) | Latex (H) | Lace (I) |
|---|---|---|---|---|---|
| 42 | ![]() | ![]() | ![]() | ![]() | ![]() |
| 123 | ![]() | ![]() | ![]() | ![]() | ![]() |
| 789 | ![]() | ![]() | ![]() | ![]() | ![]() |
All 5 materials are clearly differentiated. Patent leather has mirror gloss; velour has irregular sheen; latex has wet gloss.
Color Comparison (Red / Pink / White / Gold / Two-Tone)
| Seed | Red (J) | Pink (K) | White (L) | Gold (M) | Two-Tone (N) |
|---|---|---|---|---|---|
| 42 | ![]() | ![]() | ![]() | ![]() | ![]() |
| 123 | ![]() | ![]() | ![]() | ![]() | ![]() |
| 789 | ![]() | ![]() | ![]() | ![]() | ![]() |
All colors reproduced with high accuracy. Left-right two-tone color split also outputs accurately.
Note: All 5 materials were successfully differentiated. The clean left-right two-tone color split was an unexpected result. The tailcoat’s outcome depending on word order is worth noting as a prompt design consideration.
Key Findings
Base Keywords
bunny girlis the most stable (3/3 matching black leotard shape)playboy bunnyhas high outfit variance (white satin, bikini type, etc.)bunny girl costumeis essentially equivalent tobunny girl— no advantage from adding “costume”
Material Influence
- All 5 materials are clearly differentiated
- Satin: soft sheen, most “fabric-like” feel
- Patent Leather: mirror gloss, sharp highlights
- Velour: irregular sheen pattern, color tends toward navy to grey-purple
- Latex: wet mirror finish, high body conformity
- Lace: transparent, but higher color variance, ears tend toward fur material
Color Control
- Red, pink, white, and gold are all reproducible with high accuracy
- Two-tone color (left-right split) also outputs accurately
- White bunny outfits tend to pull pantyhose white too
Part Control
- Bunny ears: appear almost certainly (3/3 in all conditions)
- Bow tie: appears at high rate (nearly 3/3)
- Cuffs: appears at high rate (nearly 3/3)
- Tail: low appearance rate (0–1/3). Unstable even when explicitly specified
- Tailcoat jacket: doesn’t appear with
wearing tailcoat jacket over leotard. Changes toleotard under open tuxedo jacketgives 3/3 success
![[Experiment] Classic Bunny Girl Outfit Testing — Can Materials, Colors, and Parts Be Differentiated?](/tips/bunny-girl-outfit-test/A_seed123_0_0000_123.webp)
















































![[Experiment] Fashion-Style Bunny Girl Variations — How Far Can Goth, Knit, Cyber, and Shrine Maiden Share Space with Bunny Ears?](/tips/bunny-girl-casual-test/X_seed42_0_0000_42.webp)
![[Experiment] Reverse Bunny Outfit Prompt Verification — Can AI Reproduce the 'Inverted Coverage' Concept?](/tips/reverse-bunny-test/C_seed42_0_0000_42.webp)
![[Verified] Image Generation Prompt Best Practices](/tips/prompt-best-practices/cover_0_0000_4517457392071889496.webp)


