DesignNovember 13, 2026, 10:30 AM·11 min read

AI Christmas and New Year Photo: 12 Gemini Prompts and Print Sizes (2026)

12 Gemini prompts for an AI Christmas and New Year photo: portrait by the tree, pajamas on the sofa, a red-backdrop concept shoot, a Santa Claus costume, 2027 balloons, fireworks over the Bosphorus, a tombala (Turkish New Year's bingo) table, couples, babies and pets. If you plan to print the photo, we also worked out which sizes a 1K and a 2K file can handle.

AI Christmas and New Year Photo: 12 Gemini Prompts and Print Sizes (2026)
Key Takeaways
  • This post has 12 holiday prompts: portrait by the tree, a short version, pajamas on the sofa, a red-backdrop concept shoot, a Santa Claus costume, 2027 balloons, fireworks over the Bosphorus, tombala with friends, a couple decorating the tree, a baby, a pet, and turning the background of an existing photo into a New Year's scene.
  • Gemini adds things to the scene on its own: it carries over your wristwatch, puts text on the sweater, changes the color of a headscarf, and in a frame with several people it carries the street and pedestrians from the first photo into the new frame. Every prompt has a line that shuts these off.
  • If your photo shows glasses, a beard, a mole or a headscarf, write it by name in the square brackets. Delete the line for any feature you don't have; in our test, a generic 'keep the moles' line added moles to a bearded man.
  • If you plan to print, ask for a 2:3 ratio in words. The 1K file from Gemini without a plan (848x1264 pixels) gives about 212 PPI at 10x15 cm; for 20x30 cm you need the 2K file from a Google AI plan.
  • We did not test these prompts separately. The lines come from our Gemini tests between October 15 and November 12, 2026; under each prompt we note which test it comes from.

December 31 is seven weeks away. You don't need a studio booking for a New Year's photo: upload a photo where your face is clear to Gemini and describe the scene. The hard part is keeping the face the same. If you describe the tree, the light and the sweater but leave out what is on your face, Gemini fills in the gaps. Glasses can vanish, and a headscarf can change color.

That's why the 12 prompts below come in two parts: the scene on top, the face line below. We wrote the originals in Turkish, and the tests behind them used those Turkish prompts. Each one is used with your own photo, and you can paste the same text into ChatGPT along with the photo.

We did not test these 12 prompts separately

The rest of our November test budget went to cover images, and we don't have an API key to run ChatGPT's image model, so apart from the cover we ran no new tests for this post (the cover result is at the end of the post). The lines in the prompts come from our earlier tests with Nano Banana 2 Lite, Nano Banana 2 and Nano Banana Pro in the Gemini API: caricature (October 15), family photo (October 16), maternity shoot (October 21), 3D figure (October 29), newborn shoot (October 30), car (November 2), face preservation (November 3, 122 images), snow trend (November 4, 81 images), tattoo (November 6), plan comparison and hug (November 10). The people in the tests were fictional, AI-generated people.

How to Make an AI Christmas Photo in 4 Steps

  1. Open gemini.google.com or the Gemini app. You need to be at least 18 to edit a photo you upload.
  2. Use the + icon to add a waist-up photo with your face clear and good lighting. If you wear glasses or a headscarf, have them in the photo too.
  3. Copy a prompt from below, fill in the square brackets and send it in the same message as the photo. If there is more than one person, add the photos in the order you list them in the prompt.
  4. If the face comes out different, don't type 'fix it' in the same chat. Name what changed, or open a new chat and upload the photo again (see the note under Prompt 12).

12 Ready-Made Christmas Photo Prompts (Copy and Paste)

Fill in the square brackets for your own face; delete the headscarf line if it doesn't apply to you. The ratio is written in words at the end of each prompt: 4:5 for the Instagram feed, 9:16 for Stories, 2:3 for printing.

1) Portrait in front of the Christmas tree (main template)

Prompt 1: Portrait by the tree
Prompt
Create a realistic portrait of the person in the photo I uploaded, taken at home in front of a decorated Christmas tree.
[If you wear a headscarf: The headscarf stays in place, its color and the way it is tied exactly as in the photo.]
The face as in the photo: [hair color and length], [glasses, beard or mole if you have them, by name and position], same skin tone and age. Do not add moles, marks, glasses or jewelry that are not in the photo.
The person is in front of the tree, to its right (as seen by the viewer), visible from the waist up, looking at the camera. Expression: a slight closed-mouth smile, no teeth showing.
Clothing: a [color] knit sweater; no text, logo or pattern on the sweater. The watch and bracelet from the photo are not in this shot, the hands are empty.
Tree: green pine; a string of warm yellow lights; plain red and gold baubles; a gold star on top. The tree lights are slightly blurred in the background.
Light: only the warm yellow light of the tree and one lamp in the room, the face softly lit.
No text, logo or sign anywhere in the image. Only this person is in the frame.
Vertical, 4:5 aspect ratio.

The watch line comes from the November 3 test: when we moved a person into a new scene, the wristwatch came along in 19 of 32 frames, and with the line "the watch from the photo is not in this shot" it didn't appear in any of 6 frames. The expression line comes from the same test. With a short prompt that only described the scene, Nano Banana 2 turned a closed mouth into a toothy smile in 5 of 6 frames; with a prompt that spelled out the expression, it happened in 0 of 9.

2) Short version: a one-paragraph Christmas prompt

Prompt 2: Short Christmas prompt
Prompt
Place the person in this photo in front of a lit Christmas tree. Keep the face, [hair color], [glasses, beard or headscarf if any, by name] and age as in the photo; do not add moles or jewelry that are not there. Warm yellow home lighting, a plain [color] knit sweater, a slight closed-mouth smile. No text in the image, only this person in the frame. Vertical, 4:5 aspect ratio.

We kept the part that lists the face in the short version too. In our plan comparison on November 10, the gap between the model for accounts without a plan (Nano Banana 2 Lite) and the paid models closed in prompts that named the error. The short prompt has fewer of these lines, so if you don't have a plan, pick Prompt 1.

3) In pajamas on the sofa: tangerines, chestnuts and tea

Prompt 3: New Year's Eve in pajamas
Prompt
Create a realistic photo of the person in the photo I uploaded, taken at home on the sofa on New Year's Eve.
[If you wear a headscarf: The headscarf stays in place, its color and the way it is tied exactly as in the photo.]
The face as in the photo: [hair color and length], [glasses, beard or mole if you have them, by name and position], same skin tone and age. Do not add moles, marks or jewelry that are not in the photo.
The person sits cross-legged on the sofa, looking at the camera and laughing. They wear a [color] plaid cotton pajama set, with a thick cream blanket over their shoulders.
On the low wooden coffee table in front of them: five tangerines on a white plate, roasted chestnuts in a brown paper cone, and a slim, tulip-shaped tea glass with no handle on its saucer, filled with dark tea.
In the back corner, a small lit Christmas tree at waist height. The curtains are closed, the room is in warm yellow light.
No text or logo anywhere in the image. Only this person is in the frame.
Horizontal, 3:2 aspect ratio.

We wrote "no handle" on purpose. In a cover we generated on November 12, we had described the glass only as "slim, tulip-shaped", and Gemini added a handle on its own. Any detail you don't want the model to decide, write it out like this: there or not there.

4) Red backdrop studio Christmas concept

Prompt 4: Studio Christmas concept shoot
Prompt
The background is plain dark red studio paper all the way to the edges; no wrinkles, pattern or text on the backdrop.
Create a New Year's concept portrait of the person in the photo I uploaded, taken in front of this backdrop.
[If you wear a headscarf: The headscarf stays in place, its color and the way it is tied exactly as in the photo.]
The face as in the photo: [hair color and length], [glasses, beard or mole if you have them, by name and position], same skin tone and age. Do not add moles, marks, glasses or jewelry that are not in the photo.
The person sits on the floor next to three gold gift boxes. The boxes are wrapped in plain gold paper with white ribbons; no tags or text on the boxes. A few fresh pine branches on the floor.
Clothing: a [color] velvet dress, or a plain [color] shirt and dark trousers.
Light: a large, soft studio light from the left (as seen by the viewer); the backdrop is plain and shadowless.
Vertical, 4:5 aspect ratio.

We put the backdrop in the first line. In the caricature test on October 15, when we wrote the ban "do not add text, a signature or a frame" in the last line, the model added a signature and a frame in both attempts. When we described the paper in the first line as "plain all the way to the edges", it added neither. The same rule holds for a studio backdrop: describe it up front.

5) You in a Santa Claus costume

Prompt 5: Santa Claus costume
Prompt
Draw the person in the photo I uploaded in a Santa Claus costume, as a realistic photo.
The face is uncovered and recognizable as in the photo: [hair color and length], [glasses, beard or mole if you have them, by name and position], same skin tone and age. No fake white beard or mustache.
Costume (top to bottom): a red velvet hat with a white pompom, the front of the hair visible; a red velvet jacket with white fur on the collar and cuffs; a black leather belt with a plain silver buckle, no pattern or text on the buckle.
[If you wear a headscarf, use this instead of the hat: The headscarf stays in place, its color and the way it is tied exactly as in the photo; no hat on the head.]
The person stands in front of a fireplace, holding a red sack with both hands. A fire burns in the fireplace, with pine branches above it.
No text or logo anywhere in the image. Only this person is in the frame.
Vertical, 4:5 aspect ratio.

We listed every part of the costume from top to bottom. In the maternity shoot test on October 21, the piece we left out came from the input: in scenes where we didn't describe the bottoms, Nano Banana 2 carried over the jeans from the photo in 3 of 8 frames. The hat line is separate for readers who wear a headscarf. In the snow trend test on November 4, Nano Banana 2 put a beanie on a woman's head without being asked.

6) Midnight: a party with "2027" balloons

Prompt 6: 2027 balloons
Prompt
Create a realistic party photo of the person in the photo I uploaded, taken at midnight on New Year's Eve.
[If you wear a headscarf: The headscarf stays in place, its color and the way it is tied exactly as in the photo.]
The face as in the photo: [hair color and length], [glasses, beard or mole if you have them, by name and position], same skin tone and age. Do not add moles, marks or jewelry that are not in the photo.
Behind the person, at head height, four gold foil number balloons stand side by side. The balloons show exactly these four digits: "2027". No other digits, letters or text.
The person holds a lit sparkler, and the sparks light up their face. Gold and silver confetti in the air.
Clothing: a [color] sequined blouse or a plain [color] jacket.
The room is dark, with small warm yellow lights blurred in the background.
Only this person is in the frame.
Vertical, 9:16 aspect ratio.

We gave the digits in quotes, said "exactly" and shut out any other text. In the tattoo test on November 6, a name in quotes plus "copy exactly" came out right in 6 of 6 attempts, while a line listing the Turkish special letters by name turned the tested name into a misspelled one in one attempt. We did not test the balloons separately. If one of the digits comes out wrong, say "the balloons should only say 2027" in the same chat.

7) Fireworks on the Bosphorus shore in Istanbul

Prompt 7: New Year's Eve on the Bosphorus
Prompt
Create a realistic photo of the person in the photo I uploaded, taken on the shore of the Bosphorus in Istanbul on New Year's Eve.
[If you wear a headscarf: The headscarf stays in place, its color and the way it is tied exactly as in the photo.]
The face as in the photo: [hair color and length], [glasses, beard or mole if you have them, by name and position], same skin tone and age. No beanie or hat on the head, the hair is visible.
The person stands in front of the stone railing on the waterfront, waist up, facing the camera. They wear a [color] wool coat and a plain [color] scarf.
Behind them, on the far side of the water, the Bosphorus Bridge: a long suspension bridge with two tall towers, its towers and cables lit in purple and blue. Gold and red fireworks in the sky.
No one else on the shore. No signs, text or logos in the image.
Vertical, 4:5 aspect ratio.

Next to the bridge's name, we wrote what it looks like. On November 10, Nano Banana Pro turned the Süleymaniye Mosque, given only by name, into a different mosque with two minarets. In the car test on November 2, writing "bridge lights" brought up the suspension bridge over the Bosphorus in 6 of 6 frames. When you write your own city, add the color and shape next to the name of the place too.

8) Tombala (Turkish New Year's bingo) with friends (2 to 4 people)

Prompt 8: New Year's tombala
Prompt
I'm uploading [number of people] photos: photo 1 [short description], photo 2 [short description], photo 3 [short description].
Create a realistic photo of these people at home on New Year's Eve, playing tombala (a Turkish New Year's bingo game) around a round table.
Seating from left to right as seen by the viewer: the person from photo 1, the person from photo 2, the person from photo 3. Everyone is looking at the camera and laughing.
Each face as in its own photo; hair, glasses, beard, moles and headscarf in place if present. Do not add moles, marks or glasses that are not in the photos.
On the table: a red velvet tombala bag, a bowl of wooden number tokens, a plate of tangerines and mixed nuts. The tombala cards lie face down on the table, their fronts not visible.
Behind them a lit Christmas tree, the room in warm yellow light.
Only these [number of people] people are in the frame. No text in the image.
Horizontal, 3:2 aspect ratio.

We gave the positions "as seen by the viewer". In our family test on October 16, writing "to their right" put the child on the wrong side in both attempts; writing the same positions as seen by the viewer got it right in both. We asked for the cards face down: on November 4, when we asked for the sign text to be "blurry", the model drew readable but meaningless text in 4 of 4 frames, and when we described the street with "no signs and no text", no text appeared in any of 4 frames.

9) Couple: decorating the Christmas tree together

Prompt 9: Couple's Christmas photo
Prompt
I'm uploading two photos: photo 1 is me, photo 2 is [my spouse or partner].
Create a realistic photo of the two of us at home, decorating the Christmas tree together.
As seen by the viewer, on the left is the person from photo 1, holding a red bauble and hanging it on a branch. On the right, the person from photo 2 is placing a gold star on top of the tree. They look at each other and laugh.
Both faces as in their own photos: [person 1's glasses, beard, mole or headscarf details], [person 2's glasses, beard, mole or headscarf details]. Do not add moles or jewelry that are not there.
Living room: warm yellow tree lights, an open box of ornaments on the floor.
Only these two people are in the frame. No text in the image.
Vertical, 4:5 aspect ratio.

The "only these two people in the frame" line comes from the hug test on November 10. With a short prompt, both models carried the street and pedestrians from photo 1 into the new frame; once we added the location and this line, the frame came out clean. In the November 4 snow trend test, the model with the lowest face similarity in two-person frames was Lite, the model for accounts without a plan. Choose Flash or Pro in Gemini.

10) New Year's concept for a baby or child

Prompt 10: Baby New Year's concept
Prompt
Create a realistic, New Year's themed studio photo of the [age] baby in the photo I uploaded.
The face as in the photo: [hair color, or no hair], [eye color], same cheeks and face shape. Keep the same age; do not add moles or marks that are not in the photo.
The baby lies on their tummy on a thick cream wool blanket, head raised, looking at the camera. They wear a red-and-white knit romper and a matching knit beanie with a pompom.
Next to them, a small wooden sled with a short lit pine branch and two plain red baubles on it.
The background is plain light beige all the way to the edges. Soft light, gentle shadows.
No text in the image, only the baby in the frame.
Vertical, 4:5 aspect ratio.

Spell out the clothing. In the newborn test on October 30, Nano Banana Pro refused a prompt saying "the baby is naked" in one of two attempts; a prompt describing the baby covered with a blanket returned an image in 6 of 6 attempts. The person uploading the photo has to be at least 18, and before you share your child's photo, read the KVKK section below.

11) Christmas with your cat or dog

Prompt 11: Christmas with a pet
Prompt
I'm uploading two photos: photo 1 is me, photo 2 is [my cat or my dog].
Create a realistic photo of the two of us sitting on a thick rug on the floor in front of a Christmas tree.
My face as in the photo: [hair color and length], [glasses, beard, mole or headscarf if I have them, by name]. Do not add moles or jewelry that are not there.
The animal as in photo 2: [breed], [coat color], markings as seen by the viewer [where the markings are, e.g. black around the left eye]. A plain red ribbon around its neck, no hat on its head.
The animal is on my lap, looking at the camera. Plain red gift boxes under the tree, no tags.
Only me and my pet are in the frame. No text in the image.
Square, 1:1 aspect ratio.

Placing the markings "as seen by the viewer" comes from the 3D figure test on October 29: when we described the dog's markings this way, they landed in the right place in 6 of 6 frames. We have no measurement showing that animals lose a hat more easily than people do. If you want a hat, change the ribbon line to "a small red Santa hat on its head".

12) Turning the background of an existing photo into a Christmas scene

Prompt 12: Make the background Christmas
Prompt
In this photo, do not touch the person, their pose, their clothes or the [item] in their hand; change only the background.
Keep the person's face the same: [if glasses: black-framed glasses on], [if headscarf: same color and way of tying], [mole or beard if any, by name].
New background: a living room in warm yellow light, a decorated Christmas tree behind the person, the tree lights slightly blurred. Nothing from the old location in the photo remains; do not add anyone else.
Adjust the light to the face: the face lit by warm yellow light.
No text in the image. Keep the aspect ratio of the photo.

Name the glasses. In the hug test on November 10, Nano Banana 2 Lite removed glasses that weren't mentioned once we added a sentence changing the background. If the face still changes, look at the face preservation test from November 3: the message "you changed my face, fix it" left the similarity score at 0.72, while a correction that listed the features one by one raised it to an average of 0.84 over 6 frames.

If the shot comes out wrong

Look at the photo you uploaded first. If the face is small, dark or shot from the side, the model guesses the face. Start a new chat with a closer photo in good light. Don't type a one-word "fix it"; in the November 3 test, that command made the person show their teeth in 5 of 5 attempts. Name what changed: "put the glasses back on, keep the headscarf burgundy".

What Does Gemini Add to a Christmas Scene on Its Own?

Each row in the table is something Gemini added or changed without being asked in our earlier tests. The last column shows which prompts include the line that stops it.

Added on its ownTest without the lineLine that stops itWith the linePrompt
Wristwatch from the photo19 of 32 frames (Nov 3)The watch from the photo is not in this shot, hands empty0 in 6 frames1
Text and logos on the sweater or boxes8 of 20 frames (Oct 29)No text or logo in the image1 in 12 frames1-3, 5, 7-12
Headscarf changes color9 of 10 collages (Nov 4)Headscarf in place, color and tying as in the photo0 in 3 collages1, 3-7, 12
Closed mouth turns into a toothy smile5 of 6 frames (Nov 3)Mouth closed, no teeth showing0 in 9 frames1
Text asked to be blurry turns into gibberish4 of 4 frames (Nov 4)No signs or text0 in 4 frames7, 8
Street and pedestrians from the first photoin both models (Nov 10)Location + only these two people in the frameclean in both models8, 9, 11
Glasses drop offIn Pro with the word 'young' (Nov 4), in Lite with a background sentence (Nov 10)Name the glassesin place in 5 of 5 collages (Nov 4)1-9, 11, 12
Error rates in AI Christmas photo prompts drop once a line is added: wristwatch from 59% to 0, text and logos from 40% to 8%, headscarf color from 90% to 0, toothy smile from 83% to 0
How often the same error appeared in tests without and with the line (Gemini, October 29 to November 4, 2026)

The numbers are small (3 to 32 frames). They show a direction and guarantee nothing. A generic "keep the moles" line also added moles that weren't there to a bearded man in 2 of 32 frames on November 3. Delete the line entirely for any feature your face doesn't have.

Printing Your Christmas Photo: 1K or 2K?

1K is enough for 10x15 cm; 20x30 cm needs 2K. According to Google's developer documentation, Gemini's image models output 848x1264 pixels at 1K and 1696x2528 pixels at 2K in a 2:3 portrait ratio. According to the help page, in the app an account without a plan downloads 1K and a Google AI plan downloads 2K. The US print lab Nations Photo Lab asks for 500x750, 720x1080 and 1200x1800 pixels at 125, 180 and 300 PPI for a 4x6 inch (10x15 cm) print, and counts anything above 125 PPI as a good print.

Print sizeNeeded for 300 PPIWith 1K (848x1264)With 2K (1696x2528)
10x15 cm (4x6 in)1200x1800≈ 212 PPI≈ 424 PPI
20x30 cm (8x12 in)2400x3600≈ 106 PPI (below 125)≈ 212 PPI
30x45 cm (12x18 in)3600x5400≈ 71 PPI≈ 141 PPI
Christmas photo print sizes: 10x15, 20x30 and 30x45 cm print rectangles drawn to scale, each showing the PPI a Gemini 1K and 2K file gives
How many PPI Gemini's 2:3 portrait file gives at each print size (math: pixels ÷ inches)

The math: divide the pixels on the short side by the short side of the print in inches; 848 ÷ 4 = 212, 1696 ÷ 8 = 212. Ask for the ratio in words. In the snow trend test on November 4, the trend prompt asked for "2160×3840 pixels, 4K", and all 39 of 39 collages came back at 768x1376 pixels, the 1K size for 9:16; the word "vertical" made all 39 collages 9:16. For printing, end the prompt with "Vertical, 2:3 aspect ratio." A 4:5 frame printed at 10x15 gets cropped at the edges.

Gemini or ChatGPT? Making a Christmas Photo for Free

Free Gemini is enough. According to Google's help page, most image requests from accounts without a plan go to Nano Banana 2 Lite, and the file downloads at 1K. In frames with several people (Prompts 8, 9, 11), choose Flash or Pro: on November 4, Lite stayed at 0.49 to 0.55 on face similarity in a two-person frame, while Flash and Pro came out between 0.53 and 0.71.

You can use the same prompts in ChatGPT along with your photo. We did not measure ChatGPT's results or the pixel size of its downloaded file for this post. If you plan to print, read the size from the photo info on your phone and do the math above.

Common Mistakes

  • Writing only the scene. A prompt that describes the tree, the light and the sweater but never mentions the face leaves your glasses and moles up to the model.
  • Giving everyone the same generic face line. A list like "glasses, moles, headscarf the same" can make the model add things that aren't in the photo; write only what you actually have.
  • Asking for numbers and text without quotes. Instead of "balloons that say 2027", write "the balloons show exactly these four digits: "2027"".
  • Writing a city name on its own. Add what the bridge, tower or street looks like.
  • Asking for a pixel size. Writing "4K, 3840 pixels" doesn't make the file bigger; give the ratio in words, and a bigger file needs a plan.
  • Mixing up the photo order in group shots. Add the photos in the order you list them in the prompt.

You have to be at least 13 to generate images in Gemini and at least 18 to edit a photo you upload (Google help page, read on November 11, 2026). Google adds an invisible SynthID watermark to images Gemini generates, so your holiday photo carries this mark too.

Under KVKK, Turkey's personal data protection law, a photo of a face is personal data. Before you upload a friend's or partner's photo for Prompts 8 and 9, ask them. Think twice before sharing an AI photo of your child on a public account; keeping it in the family group chat or in a printed album leaves less of a trace.

Frequently Asked Questions

Yes. Upload your photo to Gemini with your Google account and paste the prompt. Without a plan, the file downloads at 1K (848x1264 pixels in 2:3 portrait), which is enough for a 10x15 cm print at about 212 PPI.

Write it in two parts: the scene (tree, light, clothing, ratio) and the face line (hair color, glasses, beard, moles or headscarf, by name). End with "no text in the image, only this person in the frame". Prompt 1 is the full version of this pattern.

For a studio look, use Prompt 4 with the red backdrop; for a home look, Prompt 1 or the pajama Prompt 3. For Stories, use Prompt 6 with the 9:16 ratio.

A fake beard and mustache cover the lower half of the face. That's what the "no fake beard or mustache" line in Prompt 5 is for. If you want the beard anyway, expect the face to look less like you.

Yes. You can paste the same prompts into ChatGPT along with your photo. The numbers in this post come from Gemini tests; we did not measure ChatGPT's results.

For 10x15 cm, the 1K file from Gemini without a plan is enough; 20x30 cm needs the 2K file from a Google AI plan. End the prompt with "Vertical, 2:3 aspect ratio" so the print isn't cropped.

We generated the cover image in Gemini (Nano Banana 2) in a single try, using an English prompt that combined the portrait from Prompt 1 with the coffee table from Prompt 3; the person is fictional. The sweater came out without text, the wrists empty, and the tea glass we described as having "no handle" has no handle. Two details missed: we asked for five tangerines and there are six on the plate, and a dining table showed up in place of the low wooden coffee table. Since we didn't specify the plate's color, it also has a red pattern on the rim. If the count or the furniture matters to you, fix them one at a time in the same chat.

We Offer This Service

Professional Photography Service

You can make the New Year's shot yourself in Gemini. When your store's holiday campaign, your menu or your product page needs a real shoot, we photograph your products, venue and team in the studio or on location and deliver the images sized for web and social media.

Design
Write to us