SOUL 2 Vs. Nano Banana Pro 💥 Higgsfield ha lanzado su nuevo generador de imágenes SOUL 2 ⚡ Puedes subirle hasta 80 imágenes de referencia de tu personaje para mantener mejor la constancia 👀 Y para compararlo bien, lo he puesto a prueba junto a Nano Banana Pro que hasta el momento es mi generador de imágenes favorito 💕 La verdad es que hay algunos resultados de SOUL 2 que me han sorprendido bastante... No está nada mal, pero sigo prefiriendo Nano Banana para la mayoría de las ocasiones 😅 Os dejo algunas imágenes que he generado y espero leer vuestras opiniones en comentarios 💌 Y si quieres los prompts de todas las imágenes comenta "ARIA" y te los mando por mensaje!

Aria Cruz | Influencer AI

@soy_aria_cruz · Digital creator

INSTAGRAM · 2026-02-26Source

1.3Klikes

920comments

Remix This

Recreate with AI Image Generator

Make your own AI viral video

Prompt

[Subject] A split-panel winter action comparison showing the same young woman rendered twice in a snowy combat-style scene. She appears in her early 20s with fair skin, a slim build, large expressive eyes, long dark hair tied into a high ponytail, round clear eyeglasses, and medium silver hoop earrings. In both panels she has an intense gritted-teeth expression and is aiming a long black firearm or tactical rifle toward the camera from behind a snowbank. Snowflakes cling to her hair and glasses. The left panel is a tighter close-up where only her face, glasses, one hand, and the gun barrel are visible over a dark foreground barrier topped with snow. The right panel is a slightly wider, more cinematic framing where she crouches behind a bright snow mound while aiming the weapon directly forward, with shoulders and winter outerwear more visible.

[Environment] Cold outdoor winter setting during active snowfall. The left panel background is muted and shallow, with an overcast gray-blue atmosphere and a dark object or barrier across the foreground. The right panel background shows a blurred snowy forest or tree line in cool blue tones. Snow should feel thick, windblown, and physically present, with flakes crossing the frame and soft accumulation on foreground surfaces.

[Composition/Camera] Two tall vertical comparison panels inside one 4:5 graphic with a dark teal outer frame and a thin central divider. The woman is the clear focal point in both halves, aligned at similar height so the comparison reads immediately. Keep the bottom labels inside each panel: a bright lime square logo with “Higgsfield SOUL 2” on the left, and a multicolor star-like logo with “NANO-BANANA PRO” on the right. The left panel should feel like a tighter, more compressed close-up; the right panel should feel slightly wider and more dramatic, but the identity must remain clearly the same person.

[Lighting] Cold diffused daylight under heavy overcast winter conditions. The light should be soft and low-contrast, with no direct sun. Skin tones remain believable but slightly cooled by the environment. Snow should read as bright but not blown out, and dark hair should still retain texture under the falling flakes.

[Style/Rendering] Photorealistic AI benchmark graphic for comparing image models on a hard winter-action prompt. The image should feel cinematic enough to be exciting, but still grounded enough that viewers can judge identity preservation, glasses consistency, snow realism, and weapon handling across both panels. Preserve believable wet snow texture, realistic facial tension, natural hoop earrings, and phone-scroll-stopping contrast between skin, dark hair, and white snow.

[Detail constraints] Keep exactly one repeated woman across two comparison panels. Preserve the glasses, hoop earrings, ponytail, snowy cover position, gritted expression, long black gun, and bottom comparison labels. Do not turn this into military battlefield chaos, fantasy combat, ski fashion, or a generic portrait in snow. No extra soldiers, explosions, helmets, camouflage squad, or dramatic fire effects. This is a focused two-panel winter action benchmark centered on the same creator identity in a snow-covered defensive pose.

Negative prompt: extra people, battlefield explosion, camouflage helmet, missing glasses, different hairstyle, outdoor sunshine, warm color grading, cartoon snow, fake weapon design, fantasy rifle, smiling expression, studio backdrop, text at top, watermark clutter, logo duplication, deformed hands, incorrect trigger grip, blurry face, anime style, oversaturated teal shadows, melted snow texture

Suggested parameters:
- Aspect ratio: 4:5
- Lens / focal length: left panel 50mm close-up feel, right panel 35mm action portrait feel
- Depth of field: moderate, face and weapon sharp, background softly blurred
- Steps: 30-40
- CFG / style strength: 6.5-8
- Sampler: DPM++ 2M Karras or equivalent
- Seed suggestion: 719334

Delta prompt strategy:
1. If the split comparison layout breaks: “two vertical winter action panels side by side with a dark teal frame and center divider”
2. If identity drifts: “same young woman repeated twice with clear round glasses, hoop earrings, high ponytail, and intense expression”
3. If the snow environment weakens: “heavy falling snow, bright snowbank foreground, cold blurred winter background”
4. If the weapon changes too much: “long black tactical rifle aimed toward camera from behind snow cover”
5. If the left panel becomes too wide: “tight close-up over snowy barrier with face and gun dominant”
6. If the right panel loses drama: “slightly wider crouched pose behind a snow mound with windblown flakes”
7. If labels disappear: “bottom panel labels for Higgsfield SOUL 2 on the left and NANO-BANANA PRO on the right”
8. If lighting becomes cinematic orange: “cold overcast blue-gray winter light, no warm sunset tones”
9. If realism drops: “photoreal snow texture, believable glasses reflections, natural skin and hair detail”
10. If the scene becomes a war movie: “single-subject benchmark portrait, no squad, no explosions, no battlefield clutter”

How soy_aria_cruz Made This Snow Battle AI Portrait and How to Recreate It

This image is effective because it compares two models on a genuinely difficult scene. Snow, motion, glasses, facial tension, weapon direction, and identity continuity all have to survive at once. That makes the post more than a cool winter frame. It becomes a stress test, and viewers can feel that immediately.

The comparison format is doing most of the strategic work. A single dramatic image might get attention, but it does not teach much. Two matched winter-action outputs with clear labels invite people to inspect differences in realism, emotion, and environmental control. That makes the image strong both as content and as evidence.

Why This Format Performs

The hardest part of the prompt is controlled aggression. The woman has to look intense without losing her recognizable face. The glasses have to remain plausible under snowfall. The gun has to aim forward in a readable way. Snow has to feel active without hiding the subject. When all of that mostly holds together, the audience experiences the result as high competence.

The second reason it works is panel contrast. The left side is tighter and more claustrophobic. The right side is wider and more cinematic. That creates a useful comparison without changing the core prompt identity. Viewers can compare not just quality, but interpretation style. One output feels more intimate; the other feels more action-oriented.

Signal	Evidence (from this image)	Mechanism	Replication Action
High-difficulty weather realism	Snowflakes, snowbanks, cold haze, and hair accumulation all need to read correctly.	Winter scenes expose whether a model can layer atmosphere without destroying the subject.	Specify both falling snow and snow-covered foreground surfaces, not only a “snowy background.”
Identity under stress	The same glasses, ponytail, earrings, and facial structure survive in both panels.	Action prompts often cause identity drift, so consistency feels impressive here.	Lock personal markers early and repeat them explicitly in the prompt.
Readable confrontation	The weapon is aimed toward the viewer in both frames.	The direct line of action creates instant tension and scroll-stopping energy.	Keep the camera relationship clear so the subject appears engaged with the frame.
Explicit benchmark packaging	The panel labels make the model showdown impossible to miss.	Viewers know exactly what they are comparing and are more likely to comment with a winner.	Use obvious but compact lower-third version labels inside the image.

Best Use Cases and Transfers

This pattern is excellent for model-vs-model posts, realism stress tests, winter action prompt packs, and comparison carousels built around difficult environmental conditions. It also transfers well to rain scenes, desert dust scenes, low-light alley scenes, and underwater portraits, because the same principle applies: keep identity stable inside a difficult atmosphere.

Best for benchmark content: the audience can judge both style and technical survival in one frame.
Best for engagement hooks: side-by-side action scenes naturally invite “left or right?” comments.
Best for prompt education: it shows that weather control and identity control are linked.
Best for advanced creator demos: the subject remains personal even inside a more cinematic scenario.

It is less suited for soft lifestyle branding, product catalog work, or emotionally calm portrait series. The image depends on tension. That is its advantage, but it narrows the kind of story it can tell.

Not ideal for comfort-content feeds: the confrontational pose changes the emotional register.
Not ideal for minimalist aesthetics: snowfall and weapon detail create deliberate visual density.
Not ideal for family-safe generic branding: the tactical framing is too aggressive.

Three Transfer Recipes

Winter realism showdown. Keep: same subject twice, heavy snowfall, lower model labels. Change: pose distance, coat design, snow density. Slot template (EN): {same creator in two winter action panels} {snow-driven atmospheric stress test} {identity preserved with glasses and hairstyle} {clear benchmark labels}
Weather-pressure portrait test. Keep: identity lock and environmental difficulty. Change: snow into rain, dust, ash, or sea spray. Slot template (EN): {same portrait subject repeated twice} {hard atmosphere partially obscuring the frame} {recognizable face markers preserved} {A/B comparison graphic}
Cinematic-versus-clean interpretation pair. Keep: one prompt, two render styles. Change: framing tension and lens feel. Slot template (EN): {left panel tighter and denser} {right panel slightly wider and more dramatic} {same subject and core action} {visible version names}

Aesthetic Read

The left panel feels harsher because the composition is compressed. The face is closer, the foreground barrier is darker, and the snow reads as interference. The right panel feels more polished because the snowbank is brighter and the background blur gives the frame cinematic depth. That difference is useful. It gives the audience something immediate to compare beyond simple sharpness.

The glasses are one of the most important details in the entire image. Without them, the subject would become a generic winter-action heroine. With them, the image still belongs to the same recognizable creator identity that appears in her softer lifestyle and sports posts. That continuity is exactly what makes the benchmark persuasive.

Observed	Why it matters for recreation
Snowflakes visible against dark hair and the air	The storm reads as active instead of decorative.
Bright snowbank foreground in both panels	The cover position gives the scene immediate action logic.
Glasses and hoop earrings remain visible	Identity markers survive even in an aggressive scene.
Close-up left versus wider right framing	The comparison shows interpretation range without changing the core concept.
Compact lower logos and model names	The benchmark packaging stays readable without overpowering the image.

Prompt Technique Breakdown

To recreate this well, you need to define the benchmark shell, then the winter-action conditions, then the identity lock. If you skip the shell, you will get one normal image. If you skip the identity lock, the two versions will feel like different women. If you skip the weather instructions, the snow becomes generic wallpaper.

Prompt chunk	What it controls	Swap ideas (EN, 2-3 options)
two vertical winter action comparison panels with a dark teal border	The A/B structure and visual packaging	split-screen benchmark graphic; side-by-side model showdown; dual-panel action comparison
same young woman with clear glasses, hoop earrings, high ponytail	Identity continuity under environmental stress	same creator repeated twice; stable facial markers; matching subject across both outputs
aiming a long black rifle from behind a snowbank with gritted teeth	The main action and emotional intensity	crouched winter confrontation; defensive aim pose; snow-cover tactical portrait
heavy falling snow and cold blurred forest background	Weather realism and cinematic atmosphere	windblown snowfall; blue-gray storm environment; active winter haze
left panel tighter, right panel slightly wider	Meaningful variation while preserving comparability	compressed close-up versus broader portrait; intimate versus cinematic framing; matched subject, different lens feel
bottom labels for Higgsfield SOUL 2 and NANO-BANANA PRO	Immediate benchmark readability	lower version tags; compact model name labels; platform logos near the bottom edge

Remix Steps

Lock the side-by-side frame first. Then stabilize the woman’s identity markers. After that, push on the snow and cover logic. The weapon and expression should come after the environment is believable, because the atmosphere is what makes the scene difficult in the first place.

Run 1: establish the split-screen benchmark composition with both subjects aligned and clearly repeated.
Run 2: lock glasses, ponytail, earrings, and facial proportions so the same person appears in both panels.
Run 3: correct the snowbanks, falling flakes, and cold background until the winter scene feels physically convincing.
Run 4: refine the aiming pose, facial tension, and lower labels to make the showdown feel intentional and readable.

If the scene becomes chaotic, simplify the background before simplifying the snow or the subject. The storm and the identity are the point. The background trees only need to support those two elements.