Skip to content
TBSI.

Start here

  • LLM LeaderboardCompare frontier models on intelligence, price and speedTool
  • What is the best superintelligence?The best frontier model for each jobModels
  • Frontier Gap ReportHow far apart the leading labs areReport
  • MethodologyHow we verify sources, data and claimsAbout

Best AI Image Prompts for Portraits and Characters (2026)

Eight prompts for cinematic, editorial, street, minimalist and fantasy portraits, with the checks to run on faces and hands

By TheBestSuperintelligence Editorial TeamUpdated

Best AI image prompts for portraits and characters

8 AI image prompts for portraits and characters to copy and adapt

Full brief
Close-up cinematic portrait of a [AGE] [GENDER] with [DISTINCTIVE FEATURE], soft dramatic lighting, shallow depth of field, 85mm lens, highly detailed skin texture, natural expression, film grain, moody atmosphere
Advanced
Editorial beauty portrait of [INVENTED SUBJECT], natural skin texture, elegant makeup, soft studio lighting, clean background, fashion magazine style, sophisticated mood
Advanced
Detailed character concept art of a [CHARACTER TYPE], full body, intricate costume design, strong silhouette, dramatic lighting, fantasy realism style, concept art quality
Advanced
Candid street style portrait of [INVENTED SUBJECT], natural light, urban background slightly blurred, authentic expression, 35mm photography style, raw and real atmosphere
Specialized
Cyberpunk character portrait, neon lights reflecting on face, intricate tech details, rain-soaked atmosphere, cinematic lighting, ultra detailed
Specialized
Epic fantasy warrior portrait, detailed armor, battle-worn face, dramatic volumetric lighting, cinematic composition, high fantasy style
Specialized
Minimalist portrait of [INVENTED SUBJECT], clean background, soft natural light, simple composition, elegant and timeless, shallow depth of field
Full brief
Powerful portrait of an elderly [MAN OR WOMAN], deep wrinkles, kind eyes, dramatic side lighting, high detail skin texture, emotional and dignified atmosphere

How to prompt AI for portraits and characters: what to add

  • The seven things Midjourney lists. Subject, medium, environment, lighting, colour, mood and composition. Add the ones that matter and leave the rest to the model.
  • What you want, not what you do not. Midjourney warns that mentioning "no cake" can still produce a cake.
  • Specific numbers. "Three people" is clearer than "people".
  • Invented subjects. The prompts use placeholders such as [INVENTED SUBJECT]. Describe traits, not a named real person.
  • Recurring characters. OpenAI documents that consistency across generations can slip, so repeat the same distinctive details and compare against your first image.
  • Camera and film terms with care. Terms like 85mm or film grain steer the look, but they are style cues, not real lens settings.

Worked example: AI image prompts for portraits, filled in

The business below is invented for this example.

Text
Powerful portrait of an elderly fisherman from a small coastal village, deep wrinkles, kind eyes, dramatic side lighting from a window, high detail skin texture, emotional and dignified atmosphere, soft grey background, head and shoulders, 4:5 vertical

A good result for this prompt:

  • Keeps the face dignified and specific, not a caricature.
  • Shows plausible hands, ears and teeth, or crops them out when they fail.
  • Uses the window light you described on one side of the face.
  • Looks like an invented person, not a recognisable real one.

Limits of this guide

These are our own untested prompts, and we did not compare their output across models. Results change between model versions. Vendors publish guidance rather than recipes. Copyright, licensing and likeness rules for generated images vary by tool and country, and are not covered here.

Sources & Further Reading

All sources fetched on 2026-11-15.

Claim Source Section or date
Describe what you want, not what you do not; use specific numbers; consider subject, medium, environment, lighting, colour, mood and composition Prompt Basics (Midjourney) Prompt Basics; read via Internet Archive copy, 2026-11-15
The image model can struggle with precise text placement, consistency of recurring characters and composition control; complex prompts may take up to 2 minutes Image generation (OpenAI) Limitations; undated

Need a different task?

Browse all 58 tasks and 345 prompts, with search and level filters.

Open the prompt library →

Frequently asked questions

What are the best AI image prompts for portraits and characters?

Name the subject, medium, setting, lighting, mood and composition, then keep the prompt short and describe what you want, not what to avoid. Use invented subjects, not real people, generate several candidates, and inspect faces, hands and teeth at full size before you use any portrait.

How do I keep a character looking the same across images?

Expect some drift. OpenAI documents that its image model may struggle to keep recurring characters or brand elements consistent across generations. Reuse the same distinctive details in every prompt, keep a log of prompts and settings, and check each new image against the first.

Is it safe to prompt for a real person's face?

Avoid it. A portrait of a named real person can mislead viewers and may raise consent and likeness issues, so use invented subjects and described traits instead. Check the licence and terms of the tool you use, because rules on use of generated images vary by tool and country.

How we sourced this. Sources were read on 2026-11-15. Thresholds and rules are as printed on each source page on that date. Methodology