The internet is buzzing with a new trend of transforming personal photos into vintage and retro-style portraits, complete with aesthetic lighting and mood. Google Gemini has been impressively accurate in generating these portraits.

We experimented with multiple prompts and even put together a guide to help create these looks. In this article, we also compare ChatGPT and Gemini to see which one produces better images and captures that Pinterest vibe.

Our experiment: Retro-style AI portraits

We experimented with generating 90’s-inspired viral retro-style AI portraits of two characters, a guy and a girl, using the same prompt across both AI platforms. The results turned out to be surprisingly different. Take a look at the prompts and the portraits created by ChatGPT and Gemini.

Results: Gemini vs. ChatGPT

Gemini delivered flawlessly crafted retro portraits that truly captured the 90’s aesthetic while keeping the Pinterest-like vibe intact. In contrast, ChatGPT altered the facial features of both the male and female characters.

Another difference was in processing time, ChatGPT took around 5–7 minutes to generate the images, whereas Gemini completed the task in under 30 seconds. That said, ChatGPT did a commendable job with the background and lighting, which aligned well with the given prompt.

FAQWhat Gemini AI does

Image Generation: Gemini can produce high-resolution images based on prompts, like vintage portraits, cinematic effects, or 3D-style figures.

Visual Styles: It can emulate specific art styles, lighting, eras (like 1940s noir or 1970s Polaroid vibes), and textures.

Direct Output: You get an actual image file you can download or post.

What ChatGPT does

Prompt Creation: It can write detailed, creative prompts that tell Gemini (or any image AI) exactly what to generate.

Image Descriptions: It can describe a scene, mood, lighting, and style in text in as much detail as needed.

ASCII or Concept Sketches: Moreover, OpenAI’s platform can produce very basic visualisations in text, but not high-quality images.

What ChatGPT cannot do

It cannot directly generate high-res images like Gemini or MidJourney. It can only guide or create the instructions for an AI that does.

So, essentially, think of it as your AI director: It crafts the vision, style, mood, and instructions. Gemini is the AI painter who brings them visually to life.

Note: Both the images used the comparison are actually AI-generated themselves.