Default AI male portraits drift toward two things: influencer-pretty Instagram faces, or generic “stock photo guy.” Neither looks like a real person. The fix is the same as for female portraits: replace empty adjectives with structural facial cues, and always specify lens, lighting, and skin texture. Ten copy-ready templates below, covering business, streetwear, athletic, cinematic, and editorial cuts.
TL;DR (June 2026): The model that renders a prompt matters as much as the prompt. For male portraits, Midjourney V8.1 (the default since April 30, 2026) gives the strongest out-of-box aesthetic; FLUX.2 and FLUX.1.1 Pro produce the most photographic skin and natural lighting; Seedream 5.0 and Nano Banana Pro (Gemini 3 Pro Image, native 4K) handle precise editing and text. The templates below are written model-agnostic in plain natural language so they paste into any of them; Midjourney-only flags like --ar and --ow are called out where they apply.
Which model to generate on (as of June 2026)
| Model | Strength for male portraits | Consistency tool | Notes |
|---|---|---|---|
| Midjourney V8.1 | Best default aesthetic, cinematic look | Omni Reference (--oref + --ow 0-100) | Default since Apr 30, 2026; 2K HD without upscale |
| FLUX.2 / FLUX.1.1 Pro | Most photographic skin & lighting | LoRA training (open weights) | Black Forest Labs; reach-for-it realism |
| Seedream 5.0 | Strong stylized + cheapest at scale | Reference image | Good for repeatable batch output |
| Nano Banana Pro | Native 4K, precise edits, readable text | Reference + conversational edits | Google Gemini 3 Pro Image |
Plain-language prompts (no flags) work in all four. The --cref flag you may have seen in older guides is the legacy V6 method and is ignored by Midjourney V7 and V8.x; use Omni Reference instead (see the consistency section).
Why default male portraits look “too pretty”
AI models lean heavily on training data biased toward symmetric, conventionally attractive faces. To break out, give it structural cues:
sharp jawline, prominent brow ridgenarrow face/square face/oval facelight stubble/full beard, well-groomed/clean-shaven- explicit eye direction (
looking directly at the lens/looking off-frame)
Drop handsome, attractive, good-looking. They don’t anchor anything and push the model back to the “influencer” mean. On Midjourney V8.1 specifically, the improved prompt adherence means a structural phrase like prominent brow ridge, slight under-eye lines now actually shows up instead of getting smoothed away, so it rewards specific bone-structure language more than V6 did.
Six-layer structure
Every realistic male portrait prompt should cover:
- Subject: age, ethnicity if relevant, face shape, hair, facial hair
- Expression: calm / brooding / focused / amused / cool
- Wardrobe: material, fit, color
- Lighting: direction, quality, temperature
- Lens: focal length, aperture, optionally film stock
- Composition: head-shot / half body / angle / background
10 copy-ready prompt templates
1. Business executive
A 32-year-old East Asian businessman, sharp jawline, slicked-back black hair, calm intelligent eyes, light stubble, tailored navy three-piece suit, silver watch, soft Rembrandt key light, plain grey backdrop, 85mm f/2.8, half body, --ar 4:52. Literary young man
A 26-year-old man, soft features, messy chestnut hair, gentle thoughtful expression, cream knit cardigan over white linen shirt, overcast soft daylight from window, 50mm f/1.8, Kodak Portra 400, holding a worn paperback, --ar 4:53. Streetwear lookbook
A 23-year-old man with shaved fade haircut, sharp brow, neutral cool expression, oversized black techwear hoodie with cyan accents, silver earrings, night street, neon cyan and magenta rim light, wet pavement reflection, 50mm f/1.4, --ar 4:54. Athletic male
A 24-year-old athletic man, toned jawline, short black hair with shaved sides, intense focused eyes, light sweat on skin, black sleeveless workout top, dramatic side rim light, deep shadow opposite side, low angle, hands wrapped in boxing tape, --ar 4:55. Cinematic male lead
A 35-year-old man with tired expressive eyes, salt-and-pepper hair, light scruff, faded leather jacket over grey crewneck, cinematic film still, anamorphic 35mm, dim practical streetlight from behind, magenta neon reflection, leaning against car, --ar 21:96. LinkedIn / corporate headshot
A 38-year-old man, clean-shaven, short side-swept dark brown hair, calm trustworthy expression, charcoal suit and white shirt, soft key light from camera-left, plain off-white backdrop, natural skin texture, 85mm f/2.8, half body, --ar 4:57. Outdoor adventurer
A 30-year-old man with sun-weathered skin, full short beard, hazel eyes, dark green waxed canvas jacket, golden hour backlight in a mountain meadow, slight wind motion in hair, 50mm f/2, Kodak Portra 800, --ar 4:58. Retro 70s editorial
A 32-year-old man with shoulder-length wavy brown hair and mustache, brown corduroy suit jacket over open collar shirt, warm tungsten lighting, slight film grain, 50mm f/2.0, Kodak Gold 200, 70s magazine aesthetic, --ar 4:59. East-Asian aesthetic / hanfu modern
A 28-year-old Chinese man in modern hanfu-inspired black robe with subtle silver embroidery, long black hair tied in a low knot, calm contemplative expression, ink-wash backdrop, soft morning light, 85mm f/2, Fuji 400H, half body, --ar 4:510. Brand ambassador (luxury)
A 40-year-old man with sharp angular features, salt-and-pepper short hair, square jaw, charcoal cashmere overcoat over white turtleneck, single beam top light, deep shadows, plain dark backdrop, magazine cover quality, 85mm f/2.8, --ar 4:5Per-use-case tuning
- Business / LinkedIn: 85mm f/2.8, Rembrandt light, grey backdrop, sober wardrobe
- Streetwear: 50mm f/1.4, neon rim light, oversized silhouette, night street
- Athletic: low angle, side rim light, sweat highlight, sleeveless / training kit
- Cinematic lead: anamorphic 35mm, leather jacket, magenta neon, contemplative gaze
- Editorial / retro: name the film stock, period-correct hairstyle, warm tungsten
Common mistakes
- Using
handsomeorattractive— empty words, model regresses to mean - No facial structure — get generic Instagram face
- No facial hair state — model picks randomly
- Specifying both
toughandsoft— contradictions = muddy result - Forgetting eye direction — gaze drifts off frame
How to keep one character consistent across shots
For a series featuring the same man, the consistency tool depends on your model (June 2026):
- Midjourney V8.1: use Omni Reference, not the old
--cref. Upload a clean front-facing portrait, then add--oref [image-url] --ow 80to your prompt.--ow(Omni Weight) runs 0 to 100; around 80 locks the face while leaving room to change wardrobe, lighting, and pose. Drop to roughly 60 if the expression feels frozen across shots. The legacy--crefflag is ignored in V7 and V8.x. - FLUX.2 / FLUX.1.1 Pro: train a small LoRA on 15-25 images of the same face for the tightest identity lock. Black Forest Labs ships open weights, so this is the most controllable route for a recurring character.
- Seedream 5.0 / Nano Banana Pro: pass a single reference image and refine conversationally (“same man, now in a charcoal coat”).
- Any model: reuse the exact facial-structure block verbatim across every prompt and only vary Style, Lighting, and Composition.
See How to Keep AI Image Style Consistent for the full workflow.
For current model pricing and parameter docs, see the official Midjourney plans page (Basic $10, Standard $30, Pro $60, Mega $120/mo as of June 2026; 20% off annual) and the Black Forest Labs FLUX site.
FAQ
Q: Which model is best for realistic male portraits in 2026? A: As of June 2026, Midjourney V8.1 (default since April 30, 2026) gives the strongest aesthetic out of the box, while FLUX.2 and FLUX.1.1 Pro from Black Forest Labs render the most photographic skin and lighting. Nano Banana Pro (Gemini 3 Pro Image) is the pick when you need native 4K or readable text in frame. The templates here paste into all of them.
Q: How do I avoid the “AI Instagram man” look?
A: Drop handsome, add sharp jawline, prominent brow ridge, narrow face, and use real film stock names. Avoid vibrant, glowing, perfect skin. On V8.1, add natural skin texture, visible pores because the model can otherwise over-smooth.
Q: Why are my male portraits always too young-looking?
A: Be specific about age. 38-year-old is better than young man. Add subtle wrinkles around eyes or salt-and-pepper hair for older subjects.
Q: How do I get convincing facial hair?
A: Describe state explicitly: light stubble, full short beard, well-groomed, mustache only, clean-shaven. Don’t leave it to the model.
Q: How do I keep the same man across multiple shots?
A: On Midjourney V8.1 use Omni Reference: add --oref [image-url] --ow 80 (the old --cref is ignored in V7/V8). On FLUX, train a LoRA on 15-25 images. Either way, reuse the same facial-structure text in every prompt.
Q: Best aspect ratio for male portraits?
A: 4:5 for social and editorial, 2:3 for full-body, 21:9 for cinematic film stills.