Gemini peut générer et traiter des images de manière conversationnelle. Vos requêtes peuvent contenir du texte, des images ou une combinaison des deux pour avoir davantage de contrôle lorsque vous créez, modifiez et itérez des visuels :
- Text-to-Image: : générez des images de haute qualité à partir de descriptions textuelles simples ou complexes.
- Image + Image-to-Text (édition) : fournissez une image et utilisez des requêtes textuelles pour ajouter, supprimer ou modifier des éléments, changer le style ou ajuster la correction colorimétrique.
- Plusieurs images vers une image (composition et transfert de style) : utilisez plusieurs images d'entrée pour composer une nouvelle scène ou transférer le style d'une image vers une autre.
- Affinement itératif : engagez une conversation pour affiner progressivement votre image en plusieurs étapes, en apportant de petites modifications jusqu'à ce qu'elle soit parfaite.
- Rendu de texte haute fidélité : générez des images précises contenant du texte lisible et bien placé, idéal pour les logos, les diagrammes et les affiches.
Toutes les images générées incluent un filigrane SynthID.
Ce guide décrit les modèles d'images Gemini 2.5 Flash (rapide) et Gemini 3 Pro Preview (avancé). Il fournit des exemples de fonctionnalités allant de la génération de texte vers image de base aux améliorations complexes et multitours, en passant par la sortie 4K et la génération ancrée dans la recherche.
Sélection du modèle
Choisissez le modèle le mieux adapté à votre cas d'utilisation spécifique.
L'aperçu d'image Gemini 3 Pro (aperçu Nano Banana Pro) est conçu pour la production d'assets professionnels et les instructions complexes. Ce modèle est ancré dans le monde réel grâce à la recherche Google. Il dispose d'un processus de "réflexion" par défaut qui affine la composition avant la génération et peut générer des images d'une résolution allant jusqu'à 4K.
Gemini 2.5 Flash Image (Nano Banana) est conçu pour la rapidité et l'efficacité. Ce modèle est optimisé pour les tâches à faible latence et à volume élevé. Il génère des images d'une résolution de 1 024 px.
Génération d'images (texte vers image)
Le code suivant montre comment générer une image à partir d'une invite descriptive.
Python
from google import genai
from google.genai import types
from PIL import Image
client = genai.Client()
prompt = (
"Create a picture of a nano banana dish in a fancy restaurant with a Gemini theme"
)
response = client.models.generate_content(
model="gemini-2.5-flash-image",
contents=[prompt],
)
for part in response.parts:
if part.text is not None:
print(part.text)
elif part.inline_data is not None:
image = part.as_image()
image.save("generated_image.png")
JavaScript
import { GoogleGenAI } from "@google/genai";
import * as fs from "node:fs";
async function main() {
const ai = new GoogleGenAI({});
const prompt =
"Create a picture of a nano banana dish in a fancy restaurant with a Gemini theme";
const response = await ai.models.generateContent({
model: "gemini-2.5-flash-image",
contents: prompt,
});
for (const part of response.candidates[0].content.parts) {
if (part.text) {
console.log(part.text);
} else if (part.inlineData) {
const imageData = part.inlineData.data;
const buffer = Buffer.from(imageData, "base64");
fs.writeFileSync("gemini-native-image.png", buffer);
console.log("Image saved as gemini-native-image.png");
}
}
}
main();
Go
package main
import (
"context"
"fmt"
"log"
"os"
"google.golang.org/genai"
)
func main() {
ctx := context.Background()
client, err := genai.NewClient(ctx, nil)
if err != nil {
log.Fatal(err)
}
result, _ := client.Models.GenerateContent(
ctx,
"gemini-2.5-flash-image",
genai.Text("Create a picture of a nano banana dish in a " +
" fancy restaurant with a Gemini theme"),
)
for _, part := range result.Candidates[0].Content.Parts {
if part.Text != "" {
fmt.Println(part.Text)
} else if part.InlineData != nil {
imageBytes := part.InlineData.Data
outputFilename := "gemini_generated_image.png"
_ = os.WriteFile(outputFilename, imageBytes, 0644)
}
}
}
Java
import com.google.genai.Client;
import com.google.genai.types.GenerateContentConfig;
import com.google.genai.types.GenerateContentResponse;
import com.google.genai.types.Part;
import java.io.IOException;
import java.nio.file.Files;
import java.nio.file.Paths;
public class TextToImage {
public static void main(String[] args) throws IOException {
try (Client client = new Client()) {
GenerateContentConfig config = GenerateContentConfig.builder()
.responseModalities("TEXT", "IMAGE")
.build();
GenerateContentResponse response = client.models.generateContent(
"gemini-3-pro-image-preview",
"Create a picture of a nano banana dish in a fancy restaurant with a Gemini theme",
config);
for (Part part : response.parts()) {
if (part.text().isPresent()) {
System.out.println(part.text().get());
} else if (part.inlineData().isPresent()) {
var blob = part.inlineData().get();
if (blob.data().isPresent()) {
Files.write(Paths.get("_01_generated_image.png"), blob.data().get());
}
}
}
}
}
}
REST
curl -s -X POST \
"https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-flash-image:generateContent" \
-H "x-goog-api-key: $GEMINI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"contents": [{
"parts": [
{"text": "Create a picture of a nano banana dish in a fancy restaurant with a Gemini theme"}
]
}]
}' \
| grep -o '"data": "[^"]*"' \
| cut -d'"' -f4 \
| base64 --decode > gemini-native-image.png
Retouche d'images (texte et image vers image)
Rappel : Assurez-vous de disposer des droits nécessaires sur toutes les images que vous importez. Ne générez aucun contenu qui porte atteinte aux droits d'autrui, y compris des vidéos ou des images trompeuses, ou qui harcèlent ou nuisent à autrui. L'utilisation de ce service d'IA générative est soumise à notre Règlement sur les utilisations interdites.
L'exemple suivant montre comment importer des images encodées en base64. Pour en savoir plus sur les images multiples, les charges utiles plus importantes et les types MIME acceptés, consultez la page Compréhension des images.
Python
from google import genai
from google.genai import types
from PIL import Image
client = genai.Client()
prompt = (
"Create a picture of my cat eating a nano-banana in a "
"fancy restaurant under the Gemini constellation",
)
image = Image.open("/path/to/cat_image.png")
response = client.models.generate_content(
model="gemini-2.5-flash-image",
contents=[prompt, image],
)
for part in response.parts:
if part.text is not None:
print(part.text)
elif part.inline_data is not None:
image = part.as_image()
image.save("generated_image.png")
JavaScript
import { GoogleGenAI } from "@google/genai";
import * as fs from "node:fs";
async function main() {
const ai = new GoogleGenAI({});
const imagePath = "path/to/cat_image.png";
const imageData = fs.readFileSync(imagePath);
const base64Image = imageData.toString("base64");
const prompt = [
{ text: "Create a picture of my cat eating a nano-banana in a" +
"fancy restaurant under the Gemini constellation" },
{
inlineData: {
mimeType: "image/png",
data: base64Image,
},
},
];
const response = await ai.models.generateContent({
model: "gemini-2.5-flash-image",
contents: prompt,
});
for (const part of response.candidates[0].content.parts) {
if (part.text) {
console.log(part.text);
} else if (part.inlineData) {
const imageData = part.inlineData.data;
const buffer = Buffer.from(imageData, "base64");
fs.writeFileSync("gemini-native-image.png", buffer);
console.log("Image saved as gemini-native-image.png");
}
}
}
main();
Go
package main
import (
"context"
"fmt"
"log"
"os"
"google.golang.org/genai"
)
func main() {
ctx := context.Background()
client, err := genai.NewClient(ctx, nil)
if err != nil {
log.Fatal(err)
}
imagePath := "/path/to/cat_image.png"
imgData, _ := os.ReadFile(imagePath)
parts := []*genai.Part{
genai.NewPartFromText("Create a picture of my cat eating a nano-banana in a fancy restaurant under the Gemini constellation"),
&genai.Part{
InlineData: &genai.Blob{
MIMEType: "image/png",
Data: imgData,
},
},
}
contents := []*genai.Content{
genai.NewContentFromParts(parts, genai.RoleUser),
}
result, _ := client.Models.GenerateContent(
ctx,
"gemini-2.5-flash-image",
contents,
)
for _, part := range result.Candidates[0].Content.Parts {
if part.Text != "" {
fmt.Println(part.Text)
} else if part.InlineData != nil {
imageBytes := part.InlineData.Data
outputFilename := "gemini_generated_image.png"
_ = os.WriteFile(outputFilename, imageBytes, 0644)
}
}
}
Java
import com.google.genai.Client;
import com.google.genai.types.Content;
import com.google.genai.types.GenerateContentConfig;
import com.google.genai.types.GenerateContentResponse;
import com.google.genai.types.Part;
import java.io.IOException;
import java.nio.file.Files;
import java.nio.file.Path;
import java.nio.file.Paths;
public class TextAndImageToImage {
public static void main(String[] args) throws IOException {
try (Client client = new Client()) {
GenerateContentConfig config = GenerateContentConfig.builder()
.responseModalities("TEXT", "IMAGE")
.build();
GenerateContentResponse response = client.models.generateContent(
"gemini-3-pro-image-preview",
Content.fromParts(
Part.fromText("""
Create a picture of my cat eating a nano-banana in
a fancy restaurant under the Gemini constellation
"""),
Part.fromBytes(
Files.readAllBytes(
Path.of("src/main/resources/cat.jpg")),
"image/jpeg")),
config);
for (Part part : response.parts()) {
if (part.text().isPresent()) {
System.out.println(part.text().get());
} else if (part.inlineData().isPresent()) {
var blob = part.inlineData().get();
if (blob.data().isPresent()) {
Files.write(Paths.get("gemini_generated_image.png"), blob.data().get());
}
}
}
}
}
}
REST
IMG_PATH=/path/to/cat_image.jpeg
if [[ "$(base64 --version 2>&1)" = *"FreeBSD"* ]]; then
B64FLAGS="--input"
else
B64FLAGS="-w0"
fi
IMG_BASE64=$(base64 "$B64FLAGS" "$IMG_PATH" 2>&1)
curl -X POST \
"https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-flash-image:generateContent" \
-H "x-goog-api-key: $GEMINI_API_KEY" \
-H 'Content-Type: application/json' \
-d "{
\"contents\": [{
\"parts\":[
{\"text\": \"'Create a picture of my cat eating a nano-banana in a fancy restaurant under the Gemini constellation\"},
{
\"inline_data\": {
\"mime_type\":\"image/jpeg\",
\"data\": \"$IMG_BASE64\"
}
}
]
}]
}" \
| grep -o '"data": "[^"]*"' \
| cut -d'"' -f4 \
| base64 --decode > gemini-edited-image.png
Modification d'images multitour
Continuez à générer et à modifier des images de manière conversationnelle. Le chat ou la conversation multitours sont la méthode recommandée pour itérer sur les images. L'exemple suivant montre une requête permettant de générer une infographie sur la photosynthèse.
Python
from google import genai
from google.genai import types
client = genai.Client()
chat = client.chats.create(
model="gemini-3-pro-image-preview",
config=types.GenerateContentConfig(
response_modalities=['TEXT', 'IMAGE'],
tools=[{"google_search": {}}]
)
)
message = "Create a vibrant infographic that explains photosynthesis as if it were a recipe for a plant's favorite food. Show the \"ingredients\" (sunlight, water, CO2) and the \"finished dish\" (sugar/energy). The style should be like a page from a colorful kids' cookbook, suitable for a 4th grader."
response = chat.send_message(message)
for part in response.parts:
if part.text is not None:
print(part.text)
elif image:= part.as_image():
image.save("photosynthesis.png")
JavaScript
import { GoogleGenAI } from "@google/genai";
const ai = new GoogleGenAI({});
async function main() {
const chat = ai.chats.create({
model: "gemini-3-pro-image-preview",
config: {
responseModalities: ['TEXT', 'IMAGE'],
tools: [{googleSearch: {}}],
},
});
await main();
const message = "Create a vibrant infographic that explains photosynthesis as if it were a recipe for a plant's favorite food. Show the \"ingredients\" (sunlight, water, CO2) and the \"finished dish\" (sugar/energy). The style should be like a page from a colorful kids' cookbook, suitable for a 4th grader."
let response = await chat.sendMessage({message});
for (const part of response.candidates[0].content.parts) {
if (part.text) {
console.log(part.text);
} else if (part.inlineData) {
const imageData = part.inlineData.data;
const buffer = Buffer.from(imageData, "base64");
fs.writeFileSync("photosynthesis.png", buffer);
console.log("Image saved as photosynthesis.png");
}
}
Go
package main
import (
"context"
"fmt"
"log"
"os"
"google.golang.org/genai"
)
func main() {
ctx := context.Background()
client, err := genai.NewClient(ctx, nil)
if err != nil {
log.Fatal(err)
}
defer client.Close()
model := client.GenerativeModel("gemini-3-pro-image-preview")
model.GenerationConfig = &pb.GenerationConfig{
ResponseModalities: []pb.ResponseModality{genai.Text, genai.Image},
}
chat := model.StartChat()
message := "Create a vibrant infographic that explains photosynthesis as if it were a recipe for a plant's favorite food. Show the \"ingredients\" (sunlight, water, CO2) and the \"finished dish\" (sugar/energy). The style should be like a page from a colorful kids' cookbook, suitable for a 4th grader."
resp, err := chat.SendMessage(ctx, genai.Text(message))
if err != nil {
log.Fatal(err)
}
for _, part := range resp.Candidates[0].Content.Parts {
if txt, ok := part.(genai.Text); ok {
fmt.Printf("%s", string(txt))
} else if img, ok := part.(genai.ImageData); ok {
err := os.WriteFile("photosynthesis.png", img.Data, 0644)
if err != nil {
log.Fatal(err)
}
}
}
}
Java
import com.google.genai.Chat;
import com.google.genai.Client;
import com.google.genai.types.Content;
import com.google.genai.types.GenerateContentConfig;
import com.google.genai.types.GenerateContentResponse;
import com.google.genai.types.GoogleSearch;
import com.google.genai.types.ImageConfig;
import com.google.genai.types.Part;
import com.google.genai.types.RetrievalConfig;
import com.google.genai.types.Tool;
import com.google.genai.types.ToolConfig;
import java.io.IOException;
import java.nio.file.Files;
import java.nio.file.Path;
import java.nio.file.Paths;
public class MultiturnImageEditing {
public static void main(String[] args) throws IOException {
try (Client client = new Client()) {
GenerateContentConfig config = GenerateContentConfig.builder()
.responseModalities("TEXT", "IMAGE")
.tools(Tool.builder()
.googleSearch(GoogleSearch.builder().build())
.build())
.build();
Chat chat = client.chats.create("gemini-3-pro-image-preview", config);
GenerateContentResponse response = chat.sendMessage("""
Create a vibrant infographic that explains photosynthesis
as if it were a recipe for a plant's favorite food.
Show the "ingredients" (sunlight, water, CO2)
and the "finished dish" (sugar/energy).
The style should be like a page from a colorful
kids' cookbook, suitable for a 4th grader.
""");
for (Part part : response.parts()) {
if (part.text().isPresent()) {
System.out.println(part.text().get());
} else if (part.inlineData().isPresent()) {
var blob = part.inlineData().get();
if (blob.data().isPresent()) {
Files.write(Paths.get("photosynthesis.png"), blob.data().get());
}
}
}
// ...
}
}
}
REST
curl -s -X POST \
"https://generativelanguage.googleapis.com/v1beta/models/gemini-3-pro-image-preview:generateContent" \
-H "x-goog-api-key: $GEMINI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"contents": [{
"role": "user",
"parts": [
{"text": "Create a vibrant infographic that explains photosynthesis as if it were a recipe for a plants favorite food. Show the \"ingredients\" (sunlight, water, CO2) and the \"finished dish\" (sugar/energy). The style should be like a page from a colorful kids cookbook, suitable for a 4th grader."}
]
}],
"generationConfig": {
"responseModalities": ["TEXT", "IMAGE"]
}
}' > turn1_response.json
cat turn1_response.json
# Requires jq to parse JSON response
jq -r '.candidates[0].content.parts[] | select(.inlineData) | .inlineData.data' turn1_response.json | head -1 | base64 --decode > photosynthesis.png
Vous pouvez ensuite utiliser la même discussion pour modifier la langue du graphique et la passer en espagnol.
Python
message = "Update this infographic to be in Spanish. Do not change any other elements of the image."
aspect_ratio = "16:9" # "1:1","2:3","3:2","3:4","4:3","4:5","5:4","9:16","16:9","21:9"
resolution = "2K" # "1K", "2K", "4K"
response = chat.send_message(message,
config=types.GenerateContentConfig(
image_config=types.ImageConfig(
aspect_ratio=aspect_ratio,
image_size=resolution
),
))
for part in response.parts:
if part.text is not None:
print(part.text)
elif image:= part.as_image():
image.save("photosynthesis_spanish.png")
JavaScript
const message = 'Update this infographic to be in Spanish. Do not change any other elements of the image.';
const aspectRatio = '16:9';
const resolution = '2K';
let response = await chat.sendMessage({
message,
config: {
responseModalities: ['TEXT', 'IMAGE'],
imageConfig: {
aspectRatio: aspectRatio,
imageSize: resolution,
},
tools: [{googleSearch: {}}],
},
});
for (const part of response.candidates[0].content.parts) {
if (part.text) {
console.log(part.text);
} else if (part.inlineData) {
const imageData = part.inlineData.data;
const buffer = Buffer.from(imageData, "base64");
fs.writeFileSync("photosynthesis2.png", buffer);
console.log("Image saved as photosynthesis2.png");
}
}
Go
message = "Update this infographic to be in Spanish. Do not change any other elements of the image."
aspect_ratio = "16:9" // "1:1","2:3","3:2","3:4","4:3","4:5","5:4","9:16","16:9","21:9"
resolution = "2K" // "1K", "2K", "4K"
model.GenerationConfig.ImageConfig = &pb.ImageConfig{
AspectRatio: aspect_ratio,
ImageSize: resolution,
}
resp, err = chat.SendMessage(ctx, genai.Text(message))
if err != nil {
log.Fatal(err)
}
for _, part := range resp.Candidates[0].Content.Parts {
if txt, ok := part.(genai.Text); ok {
fmt.Printf("%s", string(txt))
} else if img, ok := part.(genai.ImageData); ok {
err := os.WriteFile("photosynthesis_spanish.png", img.Data, 0644)
if err != nil {
log.Fatal(err)
}
}
}
Java
String aspectRatio = "16:9"; // "1:1","2:3","3:2","3:4","4:3","4:5","5:4","9:16","16:9","21:9"
String resolution = "2K"; // "1K", "2K", "4K"
config = GenerateContentConfig.builder()
.responseModalities("TEXT", "IMAGE")
.imageConfig(ImageConfig.builder()
.aspectRatio(aspectRatio)
.imageSize(resolution)
.build())
.build();
response = chat.sendMessage(
"Update this infographic to be in Spanish. " +
"Do not change any other elements of the image.",
config);
for (Part part : response.parts()) {
if (part.text().isPresent()) {
System.out.println(part.text().get());
} else if (part.inlineData().isPresent()) {
var blob = part.inlineData().get();
if (blob.data().isPresent()) {
Files.write(Paths.get("photosynthesis_spanish.png"), blob.data().get());
}
}
}
REST
# Create request2.json by combining history and new prompt
# Read model's previous response content directly into jq
jq --argjson user1 '{"role": "user", "parts": [{"text": "Create a vibrant infographic that explains photosynthesis as if it were a recipe for a plant'\''s favorite food. Show the \"ingredients\" (sunlight, water, CO2) and the \"finished dish\" (sugar/energy). The style should be like a page from a colorful kids'\'' cookbook, suitable for a 4th grader."}]}' \
--argjson user2 '{"role": "user", "parts": [{"text": "Update this infographic to be in Spanish. Do not change any other elements of the image."}]}' \
-f /dev/stdin turn1_response.json > request2.json <<'EOF_JQ_FILTER'
.candidates[0].content | {
"contents": [$user1, ., $user2],
"tools": [{"google_search": {}}],
"generationConfig": {
"responseModalities": ["TEXT", "IMAGE"],
"imageConfig": {
"aspectRatio": "16:9",
"imageSize": "2K"
}
}
}
EOF_JQ_FILTER
curl -s -X POST \
"https://generativelanguage.googleapis.com/v1beta/models/gemini-3-pro-image-preview:generateContent" \
-H "x-goog-api-key: $GEMINI_API_KEY" \
-H "Content-Type: application/json" \
-d @request2.json > turn2_response.json
jq -r '.candidates[0].content.parts[] | select(.inlineData) | .inlineData.data' turn2_response.json | head -1 | base64 --decode > photosynthesis_spanish.png
Nouveau avec Gemini 3 Pro Image
Gemini 3 Pro Image (gemini-3-pro-image-preview) est un modèle de pointe pour la génération et la retouche d'images, optimisé pour la production d'assets professionnels.
Conçu pour s'attaquer aux workflows les plus difficiles grâce à un raisonnement avancé, il excelle dans les tâches complexes de création et de modification en plusieurs étapes.
- Sortie haute résolution : fonctionnalités de génération intégrées pour les visuels 1K, 2K et 4K.
- Rendu de texte avancé : permet de générer du texte lisible et stylisé pour les infographies, les menus, les diagrammes et les supports marketing.
- Ancrage avec la recherche Google : le modèle peut utiliser la recherche Google comme outil pour vérifier des faits et générer des images basées sur des données en temps réel (par exemple, des cartes météo actuelles, des graphiques boursiers ou des événements récents).
- Mode Réflexion : le modèle utilise un processus de "réflexion" pour raisonner sur les requêtes complexes. Il génère des "images de réflexion" intermédiaires (visibles dans le backend, mais non facturées) pour affiner la composition avant de produire le résultat final de haute qualité.
- Jusqu'à 14 images de référence : vous pouvez désormais combiner jusqu'à 14 images de référence pour générer l'image finale.
Utiliser jusqu'à 14 images de référence
Gemini 3 Pro (Preview) vous permet de combiner jusqu'à 14 images de référence. Ces 14 images peuvent inclure les éléments suivants :
- Jusqu'à six images d'objets haute fidélité à inclure dans l'image finale
Jusqu'à cinq images d'êtres humains pour assurer la cohérence des personnages
Python
from google import genai
from google.genai import types
from PIL import Image
prompt = "An office group photo of these people, they are making funny faces."
aspect_ratio = "5:4" # "1:1","2:3","3:2","3:4","4:3","4:5","5:4","9:16","16:9","21:9"
resolution = "2K" # "1K", "2K", "4K"
client = genai.Client()
response = client.models.generate_content(
model="gemini-3-pro-image-preview",
contents=[
prompt,
Image.open('person1.png'),
Image.open('person2.png'),
Image.open('person3.png'),
Image.open('person4.png'),
Image.open('person5.png'),
],
config=types.GenerateContentConfig(
response_modalities=['TEXT', 'IMAGE'],
image_config=types.ImageConfig(
aspect_ratio=aspect_ratio,
image_size=resolution
),
)
)
for part in response.parts:
if part.text is not None:
print(part.text)
elif image:= part.as_image():
image.save("office.png")
JavaScript
import { GoogleGenAI } from "@google/genai";
import * as fs from "node:fs";
async function main() {
const ai = new GoogleGenAI({});
const prompt =
'An office group photo of these people, they are making funny faces.';
const aspectRatio = '5:4';
const resolution = '2K';
const contents = [
{ text: prompt },
{
inlineData: {
mimeType: "image/jpeg",
data: base64ImageFile1,
},
},
{
inlineData: {
mimeType: "image/jpeg",
data: base64ImageFile2,
},
},
{
inlineData: {
mimeType: "image/jpeg",
data: base64ImageFile3,
},
},
{
inlineData: {
mimeType: "image/jpeg",
data: base64ImageFile4,
},
},
{
inlineData: {
mimeType: "image/jpeg",
data: base64ImageFile5,
},
}
];
const response = await ai.models.generateContent({
model: 'gemini-3-pro-image-preview',
contents: contents,
config: {
responseModalities: ['TEXT', 'IMAGE'],
imageConfig: {
aspectRatio: aspectRatio,
imageSize: resolution,
},
},
});
for (const part of response.candidates[0].content.parts) {
if (part.text) {
console.log(part.text);
} else if (part.inlineData) {
const imageData = part.inlineData.data;
const buffer = Buffer.from(imageData, "base64");
fs.writeFileSync("image.png", buffer);
console.log("Image saved as image.png");
}
}
}
main();
Go
package main
import (
"context"
"fmt"
"log"
"os"
"google.golang.org/genai"
)
func main() {
ctx := context.Background()
client, err := genai.NewClient(ctx, nil)
if err != nil {
log.Fatal(err)
}
defer client.Close()
model := client.GenerativeModel("gemini-3-pro-image-preview")
model.GenerationConfig = &pb.GenerationConfig{
ResponseModalities: []pb.ResponseModality{genai.Text, genai.Image},
ImageConfig: &pb.ImageConfig{
AspectRatio: "5:4",
ImageSize: "2K",
},
}
img1, err := os.ReadFile("person1.png")
if err != nil { log.Fatal(err) }
img2, err := os.ReadFile("person2.png")
if err != nil { log.Fatal(err) }
img3, err := os.ReadFile("person3.png")
if err != nil { log.Fatal(err) }
img4, err := os.ReadFile("person4.png")
if err != nil { log.Fatal(err) }
img5, err := os.ReadFile("person5.png")
if err != nil { log.Fatal(err) }
parts := []genai.Part{
genai.Text("An office group photo of these people, they are making funny faces."),
genai.ImageData{MIMEType: "image/png", Data: img1},
genai.ImageData{MIMEType: "image/png", Data: img2},
genai.ImageData{MIMEType: "image/png", Data: img3},
genai.ImageData{MIMEType: "image/png", Data: img4},
genai.ImageData{MIMEType: "image/png", Data: img5},
}
resp, err := model.GenerateContent(ctx, parts...)
if err != nil {
log.Fatal(err)
}
for _, part := range resp.Candidates[0].Content.Parts {
if txt, ok := part.(genai.Text); ok {
fmt.Printf("%s", string(txt))
} else if img, ok := part.(genai.ImageData); ok {
err := os.WriteFile("office.png", img.Data, 0644)
if err != nil {
log.Fatal(err)
}
}
}
}
Java
import com.google.genai.Client;
import com.google.genai.types.Content;
import com.google.genai.types.GenerateContentConfig;
import com.google.genai.types.GenerateContentResponse;
import com.google.genai.types.ImageConfig;
import com.google.genai.types.Part;
import java.io.IOException;
import java.nio.file.Files;
import java.nio.file.Path;
import java.nio.file.Paths;
public class GroupPhoto {
public static void main(String[] args) throws IOException {
try (Client client = new Client()) {
GenerateContentConfig config = GenerateContentConfig.builder()
.responseModalities("TEXT", "IMAGE")
.imageConfig(ImageConfig.builder()
.aspectRatio("5:4")
.imageSize("2K")
.build())
.build();
GenerateContentResponse response = client.models.generateContent(
"gemini-3-pro-image-preview",
Content.fromParts(
Part.fromText("An office group photo of these people, they are making funny faces."),
Part.fromBytes(Files.readAllBytes(Path.of("person1.png")), "image/png"),
Part.fromBytes(Files.readAllBytes(Path.of("person2.png")), "image/png"),
Part.fromBytes(Files.readAllBytes(Path.of("person3.png")), "image/png"),
Part.fromBytes(Files.readAllBytes(Path.of("person4.png")), "image/png"),
Part.fromBytes(Files.readAllBytes(Path.of("person5.png")), "image/png")
), config);
for (Part part : response.parts()) {
if (part.text().isPresent()) {
System.out.println(part.text().get());
} else if (part.inlineData().isPresent()) {
var blob = part.inlineData().get();
if (blob.data().isPresent()) {
Files.write(Paths.get("office.png"), blob.data().get());
}
}
}
}
}
}
REST
IMG_PATH1=person1.png
IMG_PATH2=person2.png
IMG_PATH3=person3.png
IMG_PATH4=person4.png
IMG_PATH5=person5.png
if [[ "$(base64 --version 2>&1)" = *"FreeBSD"* ]]; then
B64FLAGS="--input"
else
B64FLAGS="-w0"
fi
IMG1_BASE64=$(base64 "$B64FLAGS" "$IMG_PATH1" 2>&1)
IMG2_BASE64=$(base64 "$B64FLAGS" "$IMG_PATH2" 2>&1)
IMG3_BASE64=$(base64 "$B64FLAGS" "$IMG_PATH3" 2>&1)
IMG4_BASE64=$(base64 "$B64FLAGS" "$IMG_PATH4" 2>&1)
IMG5_BASE64=$(base64 "$B64FLAGS" "$IMG_PATH5" 2>&1)
curl -X POST \
"https://generativelanguage.googleapis.com/v1beta/models/gemini-3-pro-image-preview:generateContent" \
-H "x-goog-api-key: $GEMINI_API_KEY" \
-H 'Content-Type: application/json' \
-d "{
\"contents\": [{
\"parts\":[
{\"text\": \"An office group photo of these people, they are making funny faces.\"},
{\"inline_data\": {\"mime_type\":\"image/png\", \"data\": \"$IMG1_BASE64\"}},
{\"inline_data\": {\"mime_type\":\"image/png\", \"data\": \"$IMG2_BASE64\"}},
{\"inline_data\": {\"mime_type\":\"image/png\", \"data\": \"$IMG3_BASE64\"}},
{\"inline_data\": {\"mime_type\":\"image/png\", \"data\": \"$IMG4_BASE64\"}},
{\"inline_data\": {\"mime_type\":\"image/png\", \"data\": \"$IMG5_BASE64\"}}
]
}],
\"generationConfig\": {
\"responseModalities\": [\"TEXT\", \"IMAGE\"],
\"imageConfig\": {
\"aspectRatio\": \"5:4\",
\"imageSize\": \"2K\"
}
}
}" | jq -r '.candidates[0].content.parts[] | select(.inlineData) | .inlineData.data' | head -1 | base64 --decode > office.png
Ancrage avec la recherche Google
Utilisez l'outil de recherche Google pour générer des images basées sur des informations en temps réel, comme les prévisions météo, les graphiques boursiers ou les événements récents.
Remarques à prendre en compte lorsque vous utilisez l'ancrage avec la recherche Google pour générer des images :
- Les résultats de recherche basés sur des images ne sont pas transmis au modèle de génération et sont exclus de la réponse.
Le mode Image uniquement (
responseModalities = ["IMAGE"]) ne renvoie pas de résultat d'image lorsqu'il est utilisé avec l'ancrage avec la recherche Google.
Python
from google import genai
prompt = "Visualize the current weather forecast for the next 5 days in San Francisco as a clean, modern weather chart. Add a visual on what I should wear each day"
aspect_ratio = "16:9" # "1:1","2:3","3:2","3:4","4:3","4:5","5:4","9:16","16:9","21:9"
client = genai.Client()
response = client.models.generate_content(
model="gemini-3-pro-image-preview",
contents=prompt,
config=types.GenerateContentConfig(
response_modalities=['Text', 'Image'],
image_config=types.ImageConfig(
aspect_ratio=aspect_ratio,
),
tools=[{"google_search": {}}]
)
)
for part in response.parts:
if part.text is not None:
print(part.text)
elif image:= part.as_image():
image.save("weather.png")
JavaScript
import { GoogleGenAI } from "@google/genai";
import * as fs from "node:fs";
async function main() {
const ai = new GoogleGenAI({});
const prompt = 'Visualize the current weather forecast for the next 5 days in San Francisco as a clean, modern weather chart. Add a visual on what I should wear each day';
const aspectRatio = '16:9';
const resolution = '2K';
const response = await ai.models.generateContent({
model: 'gemini-3-pro-image-preview',
contents: prompt,
config: {
responseModalities: ['TEXT', 'IMAGE'],
imageConfig: {
aspectRatio: aspectRatio,
imageSize: resolution,
},
},
});
for (const part of response.candidates[0].content.parts) {
if (part.text) {
console.log(part.text);
} else if (part.inlineData) {
const imageData = part.inlineData.data;
const buffer = Buffer.from(imageData, "base64");
fs.writeFileSync("image.png", buffer);
console.log("Image saved as image.png");
}
}
}
main();
Java
import com.google.genai.Client;
import com.google.genai.types.GenerateContentConfig;
import com.google.genai.types.GenerateContentResponse;
import com.google.genai.types.GoogleSearch;
import com.google.genai.types.ImageConfig;
import com.google.genai.types.Part;
import com.google.genai.types.Tool;
import java.io.IOException;
import java.nio.file.Files;
import java.nio.file.Paths;
public class SearchGrounding {
public static void main(String[] args) throws IOException {
try (Client client = new Client()) {
GenerateContentConfig config = GenerateContentConfig.builder()
.responseModalities("TEXT", "IMAGE")
.imageConfig(ImageConfig.builder()
.aspectRatio("16:9")
.build())
.tools(Tool.builder()
.googleSearch(GoogleSearch.builder().build())
.build())
.build();
GenerateContentResponse response = client.models.generateContent(
"gemini-3-pro-image-preview", """
Visualize the current weather forecast for the next 5 days
in San Francisco as a clean, modern weather chart.
Add a visual on what I should wear each day
""",
config);
for (Part part : response.parts()) {
if (part.text().isPresent()) {
System.out.println(part.text().get());
} else if (part.inlineData().isPresent()) {
var blob = part.inlineData().get();
if (blob.data().isPresent()) {
Files.write(Paths.get("weather.png"), blob.data().get());
}
}
}
}
}
}
REST
curl -s -X POST \
"https://generativelanguage.googleapis.com/v1beta/models/gemini-3-pro-image-preview:generateContent" \
-H "x-goog-api-key: $GEMINI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"contents": [{"parts": [{"text": "Visualize the current weather forecast for the next 5 days in San Francisco as a clean, modern weather chart. Add a visual on what I should wear each day"}]}],
"tools": [{"google_search": {}}],
"generationConfig": {
"responseModalities": ["TEXT", "IMAGE"],
"imageConfig": {"aspectRatio": "16:9"}
}
}' | jq -r '.candidates[0].content.parts[] | select(.inlineData) | .inlineData.data' | head -1 | base64 --decode > weather.png
La réponse inclut groundingMetadata, qui contient les champs obligatoires suivants :
searchEntryPoint: contient le code HTML et CSS permettant d'afficher les suggestions de recherche requises.groundingChunks: renvoie les trois principales sources Web utilisées pour ancrer l'image générée.
Générer des images jusqu'à la résolution 4K
Gemini 3 Pro Image génère par défaut des images de 1 K, mais peut également générer des images de 2 K et 4 K. Pour générer des composants de résolution supérieure, spécifiez image_size dans generation_config.
Vous devez utiliser un "K" majuscule (par exemple, 1K, 2K, 4K). Paramètres en minuscules (par exemple, 1 000) seront refusées.
Python
from google import genai
from google.genai import types
prompt = "Da Vinci style anatomical sketch of a dissected Monarch butterfly. Detailed drawings of the head, wings, and legs on textured parchment with notes in English."
aspect_ratio = "1:1" # "1:1","2:3","3:2","3:4","4:3","4:5","5:4","9:16","16:9","21:9"
resolution = "1K" # "1K", "2K", "4K"
client = genai.Client()
response = client.models.generate_content(
model="gemini-3-pro-image-preview",
contents=prompt,
config=types.GenerateContentConfig(
response_modalities=['TEXT', 'IMAGE'],
image_config=types.ImageConfig(
aspect_ratio=aspect_ratio,
image_size=resolution
),
)
)
for part in response.parts:
if part.text is not None:
print(part.text)
elif image:= part.as_image():
image.save("butterfly.png")
JavaScript
import { GoogleGenAI } from "@google/genai";
import * as fs from "node:fs";
async function main() {
const ai = new GoogleGenAI({});
const prompt =
'Da Vinci style anatomical sketch of a dissected Monarch butterfly. Detailed drawings of the head, wings, and legs on textured parchment with notes in English.';
const aspectRatio = '1:1';
const resolution = '1K';
const response = await ai.models.generateContent({
model: 'gemini-3-pro-image-preview',
contents: prompt,
config: {
responseModalities: ['TEXT', 'IMAGE'],
imageConfig: {
aspectRatio: aspectRatio,
imageSize: resolution,
},
},
});
for (const part of response.candidates[0].content.parts) {
if (part.text) {
console.log(part.text);
} else if (part.inlineData) {
const imageData = part.inlineData.data;
const buffer = Buffer.from(imageData, "base64");
fs.writeFileSync("image.png", buffer);
console.log("Image saved as image.png");
}
}
}
main();
Go
package main
import (
"context"
"fmt"
"log"
"os"
"google.golang.org/genai"
)
func main() {
ctx := context.Background()
client, err := genai.NewClient(ctx, nil)
if err != nil {
log.Fatal(err)
}
defer client.Close()
model := client.GenerativeModel("gemini-3-pro-image-preview")
model.GenerationConfig = &pb.GenerationConfig{
ResponseModalities: []pb.ResponseModality{genai.Text, genai.Image},
ImageConfig: &pb.ImageConfig{
AspectRatio: "1:1",
ImageSize: "1K",
},
}
prompt := "Da Vinci style anatomical sketch of a dissected Monarch butterfly. Detailed drawings of the head, wings, and legs on textured parchment with notes in English."
resp, err := model.GenerateContent(ctx, genai.Text(prompt))
if err != nil {
log.Fatal(err)
}
for _, part := range resp.Candidates[0].Content.Parts {
if txt, ok := part.(genai.Text); ok {
fmt.Printf("%s", string(txt))
} else if img, ok := part.(genai.ImageData); ok {
err := os.WriteFile("butterfly.png", img.Data, 0644)
if err != nil {
log.Fatal(err)
}
}
}
}
Java
import com.google.genai.Client;
import com.google.genai.types.GenerateContentConfig;
import com.google.genai.types.GenerateContentResponse;
import com.google.genai.types.GoogleSearch;
import com.google.genai.types.ImageConfig;
import com.google.genai.types.Part;
import com.google.genai.types.Tool;
import java.io.IOException;
import java.nio.file.Files;
import java.nio.file.Paths;
public class HiRes {
public static void main(String[] args) throws IOException {
try (Client client = new Client()) {
GenerateContentConfig config = GenerateContentConfig.builder()
.responseModalities("TEXT", "IMAGE")
.imageConfig(ImageConfig.builder()
.aspectRatio("16:9")
.imageSize("4K")
.build())
.build();
GenerateContentResponse response = client.models.generateContent(
"gemini-3-pro-image-preview", """
Da Vinci style anatomical sketch of a dissected Monarch butterfly.
Detailed drawings of the head, wings, and legs on textured
parchment with notes in English.
""",
config);
for (Part part : response.parts()) {
if (part.text().isPresent()) {
System.out.println(part.text().get());
} else if (part.inlineData().isPresent()) {
var blob = part.inlineData().get();
if (blob.data().isPresent()) {
Files.write(Paths.get("butterfly.png"), blob.data().get());
}
}
}
}
}
}
REST
curl -s -X POST \
"https://generativelanguage.googleapis.com/v1beta/models/gemini-3-pro-image-preview:generateContent" \
-H "x-goog-api-key: $GEMINI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"contents": [{"parts": [{"text": "Da Vinci style anatomical sketch of a dissected Monarch butterfly. Detailed drawings of the head, wings, and legs on textured parchment with notes in English."}]}],
"tools": [{"google_search": {}}],
"generationConfig": {
"responseModalities": ["TEXT", "IMAGE"],
"imageConfig": {"aspectRatio": "1:1", "imageSize": "1K"}
}
}' | jq -r '.candidates[0].content.parts[] | select(.inlineData) | .inlineData.data' | head -1 | base64 --decode > butterfly.png
Voici un exemple d'image générée à partir de cette requête :
Processus de réflexion
Le modèle Gemini 3 Pro Image Preview est un modèle de réflexion qui utilise un processus de raisonnement ("Réflexion") pour les requêtes complexes. Cette fonctionnalité est activée par défaut et ne peut pas être désactivée dans l'API. Pour en savoir plus sur le processus de réflexion, consultez le guide Réflexion de Gemini.
Le modèle génère jusqu'à deux images intermédiaires pour tester la composition et la logique. La dernière image de la section "Réflexion" est également l'image rendue finale.
Vous pouvez consulter les réflexions qui ont conduit à la production de l'image finale.
Python
for part in response.parts:
if part.thought:
if part.text:
print(part.text)
elif image:= part.as_image():
image.show()
JavaScript
for (const part of response.candidates[0].content.parts) {
if (part.thought) {
if (part.text) {
console.log(part.text);
} else if (part.inlineData) {
const imageData = part.inlineData.data;
const buffer = Buffer.from(imageData, 'base64');
fs.writeFileSync('image.png', buffer);
console.log('Image saved as image.png');
}
}
}
Signatures de pensée
Les signatures de pensée sont des représentations chiffrées du processus de réflexion interne du modèle. Elles sont utilisées pour préserver le contexte de raisonnement lors des interactions multitours. Toutes les réponses incluent un champ thought_signature. En règle générale, si vous recevez une signature de pensée dans une réponse du modèle, vous devez la renvoyer exactement telle qu'elle a été reçue lorsque vous envoyez l'historique des conversations au tour suivant. Si les signatures de pensée ne sont pas diffusées, la réponse peut échouer. Consultez la documentation sur la signature de pensée pour en savoir plus sur les signatures en général.
Voici comment fonctionnent les signatures de pensée :
- Toutes les parties
inline_dataavec l'imagemimetypequi font partie de la réponse doivent comporter une signature. - S'il y a des parties de texte au début (avant toute image) juste après les pensées, la première partie de texte doit également avoir une signature.
- Les réflexions n'ont pas de signatures. Si des parties
inline_dataavec l'imagemimetypefont partie des réflexions, elles n'auront pas de signatures.
Le code suivant montre un exemple d'emplacement des signatures de pensée :
[
{
"inline_data": {
"data": "<base64_image_data_0>",
"mime_type": "image/png"
},
"thought": true // Thoughts don't have signatures
},
{
"inline_data": {
"data": "<base64_image_data_1>",
"mime_type": "image/png"
},
"thought": true // Thoughts don't have signatures
},
{
"inline_data": {
"data": "<base64_image_data_2>",
"mime_type": "image/png"
},
"thought": true // Thoughts don't have signatures
},
{
"text": "Here is a step-by-step guide to baking macarons, presented in three separate images.\n\n### Step 1: Piping the Batter\n\nThe first step after making your macaron batter is to pipe it onto a baking sheet. This requires a steady hand to create uniform circles.\n\n",
"thought_signature": "<Signature_A>" // The first non-thought part always has a signature
},
{
"inline_data": {
"data": "<base64_image_data_3>",
"mime_type": "image/png"
},
"thought_signature": "<Signature_B>" // All image parts have a signatures
},
{
"text": "\n\n### Step 2: Baking and Developing Feet\n\nOnce piped, the macarons are baked in the oven. A key sign of a successful bake is the development of \"feet\"—the ruffled edge at the base of each macaron shell.\n\n"
// Follow-up text parts don't have signatures
},
{
"inline_data": {
"data": "<base64_image_data_4>",
"mime_type": "image/png"
},
"thought_signature": "<Signature_C>" // All image parts have a signatures
},
{
"text": "\n\n### Step 3: Assembling the Macaron\n\nThe final step is to pair the cooled macaron shells by size and sandwich them together with your desired filling, creating the classic macaron dessert.\n\n"
},
{
"inline_data": {
"data": "<base64_image_data_5>",
"mime_type": "image/png"
},
"thought_signature": "<Signature_D>" // All image parts have a signatures
}
]
Autres modes de génération d'images
Gemini prend en charge d'autres modes d'interaction avec les images en fonction de la structure et du contexte de la requête, y compris :
- Texte vers image(s) et texte (entrelacé) : génère des images avec du texte associé.
- Exemple de requête : "Génère une recette illustrée de paella."
- Image(s) et texte vers image(s) et texte (entrelacés) : utilise des images et du texte en entrée pour créer de nouvelles images et du texte associés.
- Exemple de requête : (avec l'image d'une pièce meublée) "Quelles autres couleurs de canapés conviendraient à mon espace ? Peux-tu modifier l'image ?"
Guide et stratégies de requête
Pour maîtriser la génération d'images, vous devez commencer par un principe fondamental :
Décrivez la scène, ne vous contentez pas de lister des mots clés. La principale force du modèle réside dans sa compréhension approfondie du langage. Un paragraphe descriptif et narratif produira presque toujours une image de meilleure qualité et plus cohérente qu'une liste de mots isolés.
Requêtes pour générer des images
Les stratégies suivantes vous aideront à créer des requêtes efficaces pour générer exactement les images que vous recherchez.
1. Scènes photoréalistes
Pour obtenir des images réalistes, utilisez des termes de photographie. Mentionnez les angles de caméra, les types d'objectifs, l'éclairage et les détails pour guider le modèle vers un résultat photoréaliste.
Modèle
A photorealistic [shot type] of [subject], [action or expression], set in
[environment]. The scene is illuminated by [lighting description], creating
a [mood] atmosphere. Captured with a [camera/lens details], emphasizing
[key textures and details]. The image should be in a [aspect ratio] format.
Prompt
A photorealistic close-up portrait of an elderly Japanese ceramicist with
deep, sun-etched wrinkles and a warm, knowing smile. He is carefully
inspecting a freshly glazed tea bowl. The setting is his rustic,
sun-drenched workshop. The scene is illuminated by soft, golden hour light
streaming through a window, highlighting the fine texture of the clay.
Captured with an 85mm portrait lens, resulting in a soft, blurred background
(bokeh). The overall mood is serene and masterful. Vertical portrait
orientation.
Python
from google import genai
from google.genai import types
client = genai.Client()
response = client.models.generate_content(
model="gemini-2.5-flash-image",
contents="A photorealistic close-up portrait of an elderly Japanese ceramicist with deep, sun-etched wrinkles and a warm, knowing smile. He is carefully inspecting a freshly glazed tea bowl. The setting is his rustic, sun-drenched workshop with pottery wheels and shelves of clay pots in the background. The scene is illuminated by soft, golden hour light streaming through a window, highlighting the fine texture of the clay and the fabric of his apron. Captured with an 85mm portrait lens, resulting in a soft, blurred background (bokeh). The overall mood is serene and masterful.",
)
for part in response.parts:
if part.text is not None:
print(part.text)
elif part.inline_data is not None:
image = part.as_image()
image.save("photorealistic_example.png")
JavaScript
import { GoogleGenAI } from "@google/genai";
import * as fs from "node:fs";
async function main() {
const ai = new GoogleGenAI({});
const prompt =
"A photorealistic close-up portrait of an elderly Japanese ceramicist with deep, sun-etched wrinkles and a warm, knowing smile. He is carefully inspecting a freshly glazed tea bowl. The setting is his rustic, sun-drenched workshop with pottery wheels and shelves of clay pots in the background. The scene is illuminated by soft, golden hour light streaming through a window, highlighting the fine texture of the clay and the fabric of his apron. Captured with an 85mm portrait lens, resulting in a soft, blurred background (bokeh). The overall mood is serene and masterful.";
const response = await ai.models.generateContent({
model: "gemini-2.5-flash-image",
contents: prompt,
});
for (const part of response.candidates[0].content.parts) {
if (part.text) {
console.log(part.text);
} else if (part.inlineData) {
const imageData = part.inlineData.data;
const buffer = Buffer.from(imageData, "base64");
fs.writeFileSync("photorealistic_example.png", buffer);
console.log("Image saved as photorealistic_example.png");
}
}
}
main();
Go
package main
import (
"context"
"fmt"
"log"
"os"
"google.golang.org/genai"
)
func main() {
ctx := context.Background()
client, err := genai.NewClient(ctx, nil)
if err != nil {
log.Fatal(err)
}
result, _ := client.Models.GenerateContent(
ctx,
"gemini-2.5-flash-image",
genai.Text("A photorealistic close-up portrait of an elderly Japanese ceramicist with deep, sun-etched wrinkles and a warm, knowing smile. He is carefully inspecting a freshly glazed tea bowl. The setting is his rustic, sun-drenched workshop with pottery wheels and shelves of clay pots in the background. The scene is illuminated by soft, golden hour light streaming through a window, highlighting the fine texture of the clay and the fabric of his apron. Captured with an 85mm portrait lens, resulting in a soft, blurred background (bokeh). The overall mood is serene and masterful."),
)
for _, part := range result.Candidates[0].Content.Parts {
if part.Text != "" {
fmt.Println(part.Text)
} else if part.InlineData != nil {
imageBytes := part.InlineData.Data
outputFilename := "photorealistic_example.png"
_ = os.WriteFile(outputFilename, imageBytes, 0644)
}
}
}
REST
curl -s -X POST
"https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-flash-image:generateContent" \
-H "x-goog-api-key: $GEMINI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"contents": [{
"parts": [
{"text": "A photorealistic close-up portrait of an elderly Japanese ceramicist with deep, sun-etched wrinkles and a warm, knowing smile. He is carefully inspecting a freshly glazed tea bowl. The setting is his rustic, sun-drenched workshop with pottery wheels and shelves of clay pots in the background. The scene is illuminated by soft, golden hour light streaming through a window, highlighting the fine texture of the clay and the fabric of his apron. Captured with an 85mm portrait lens, resulting in a soft, blurred background (bokeh). The overall mood is serene and masterful."}
]
}]
}' \
| grep -o '"data": "[^"]*"' \
| cut -d'"' -f4 \
| base64 --decode > photorealistic_example.png
2. Illustrations et autocollants stylisés
Pour créer des autocollants, des icônes ou des composants, soyez explicite sur le style et demandez un arrière-plan transparent.
Modèle
A [style] sticker of a [subject], featuring [key characteristics] and a
[color palette]. The design should have [line style] and [shading style].
The background must be transparent.
Prompt
A kawaii-style sticker of a happy red panda wearing a tiny bamboo hat. It's
munching on a green bamboo leaf. The design features bold, clean outlines,
simple cel-shading, and a vibrant color palette. The background must be white.
Python
from google import genai
from google.genai import types
client = genai.Client()
response = client.models.generate_content(
model="gemini-2.5-flash-image",
contents="A kawaii-style sticker of a happy red panda wearing a tiny bamboo hat. It's munching on a green bamboo leaf. The design features bold, clean outlines, simple cel-shading, and a vibrant color palette. The background must be white.",
)
for part in response.parts:
if part.text is not None:
print(part.text)
elif part.inline_data is not None:
image = part.as_image()
image.save("red_panda_sticker.png")
JavaScript
import { GoogleGenAI } from "@google/genai";
import * as fs from "node:fs";
async function main() {
const ai = new GoogleGenAI({});
const prompt =
"A kawaii-style sticker of a happy red panda wearing a tiny bamboo hat. It's munching on a green bamboo leaf. The design features bold, clean outlines, simple cel-shading, and a vibrant color palette. The background must be white.";
const response = await ai.models.generateContent({
model: "gemini-2.5-flash-image",
contents: prompt,
});
for (const part of response.candidates[0].content.parts) {
if (part.text) {
console.log(part.text);
} else if (part.inlineData) {
const imageData = part.inlineData.data;
const buffer = Buffer.from(imageData, "base64");
fs.writeFileSync("red_panda_sticker.png", buffer);
console.log("Image saved as red_panda_sticker.png");
}
}
}
main();
Go
package main
import (
"context"
"fmt"
"log"
"os"
"google.golang.org/genai"
)
func main() {
ctx := context.Background()
client, err := genai.NewClient(ctx, nil)
if err != nil {
log.Fatal(err)
}
result, _ := client.Models.GenerateContent(
ctx,
"gemini-2.5-flash-image",
genai.Text("A kawaii-style sticker of a happy red panda wearing a tiny bamboo hat. It's munching on a green bamboo leaf. The design features bold, clean outlines, simple cel-shading, and a vibrant color palette. The background must be white."),
)
for _, part := range result.Candidates[0].Content.Parts {
if part.Text != "" {
fmt.Println(part.Text)
} else if part.InlineData != nil {
imageBytes := part.InlineData.Data
outputFilename := "red_panda_sticker.png"
_ = os.WriteFile(outputFilename, imageBytes, 0644)
}
}
}
REST
curl -s -X POST
"https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-flash-image:generateContent" \
-H "x-goog-api-key: $GEMINI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"contents": [{
"parts": [
{"text": "A kawaii-style sticker of a happy red panda wearing a tiny bamboo hat. It'"'"'s munching on a green bamboo leaf. The design features bold, clean outlines, simple cel-shading, and a vibrant color palette. The background must be white."}
]
}]
}' \
| grep -o '"data": "[^"]*"' \
| cut -d'"' -f4 \
| base64 --decode > red_panda_sticker.png
3. Texte exact dans les images
Gemini excelle dans le rendu de texte. Décrivez clairement le texte, le style de police et la conception globale.
Modèle
Create a [image type] for [brand/concept] with the text "[text to render]"
in a [font style]. The design should be [style description], with a
[color scheme].
Prompt
Create a modern, minimalist logo for a coffee shop called 'The Daily Grind'.
The text should be in a clean, bold, sans-serif font. The design should
feature a simple, stylized icon of a a coffee bean seamlessly integrated
with the text. The color scheme is black and white.
Python
from google import genai
from google.genai import types
client = genai.Client()
response = client.models.generate_content(
model="gemini-2.5-flash-image",
contents="Create a modern, minimalist logo for a coffee shop called 'The Daily Grind'. The text should be in a clean, bold, sans-serif font. The design should feature a simple, stylized icon of a a coffee bean seamlessly integrated with the text. The color scheme is black and white.",
)
for part in response.parts:
if part.text is not None:
print(part.text)
elif part.inline_data is not None:
image = part.as_image()
image.save("logo_example.png")
JavaScript
import { GoogleGenAI } from "@google/genai";
import * as fs from "node:fs";
async function main() {
const ai = new GoogleGenAI({});
const prompt =
"Create a modern, minimalist logo for a coffee shop called 'The Daily Grind'. The text should be in a clean, bold, sans-serif font. The design should feature a simple, stylized icon of a a coffee bean seamlessly integrated with the text. The color scheme is black and white.";
const response = await ai.models.generateContent({
model: "gemini-2.5-flash-image",
contents: prompt,
});
for (const part of response.candidates[0].content.parts) {
if (part.text) {
console.log(part.text);
} else if (part.inlineData) {
const imageData = part.inlineData.data;
const buffer = Buffer.from(imageData, "base64");
fs.writeFileSync("logo_example.png", buffer);
console.log("Image saved as logo_example.png");
}
}
}
main();
Go
package main
import (
"context"
"fmt"
"log"
"os"
"google.golang.org/genai"
)
func main() {
ctx := context.Background()
client, err := genai.NewClient(ctx, nil)
if err != nil {
log.Fatal(err)
}
result, _ := client.Models.GenerateContent(
ctx,
"gemini-2.5-flash-image",
genai.Text("Create a modern, minimalist logo for a coffee shop called 'The Daily Grind'. The text should be in a clean, bold, sans-serif font. The design should feature a simple, stylized icon of a a coffee bean seamlessly integrated with the text. The color scheme is black and white."),
)
for _, part := range result.Candidates[0].Content.Parts {
if part.Text != "" {
fmt.Println(part.Text)
} else if part.InlineData != nil {
imageBytes := part.InlineData.Data
outputFilename := "logo_example.png"
_ = os.WriteFile(outputFilename, imageBytes, 0644)
}
}
}
REST
curl -s -X POST
"https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-flash-image:generateContent" \
-H "x-goog-api-key: $GEMINI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"contents": [{
"parts": [
{"text": "Create a modern, minimalist logo for a coffee shop called '"'"'The Daily Grind'"'"'. The text should be in a clean, bold, sans-serif font. The design should feature a simple, stylized icon of a a coffee bean seamlessly integrated with the text. The color scheme is black and white."}
]
}]
}' \
| grep -o '"data": "[^"]*"' \
| cut -d'"' -f4 \
| base64 --decode > logo_example.png
4. Maquettes de produits et photographie commerciale
Idéal pour créer des photos de produits claires et professionnelles pour l'e-commerce, la publicité ou le branding.
Modèle
A high-resolution, studio-lit product photograph of a [product description]
on a [background surface/description]. The lighting is a [lighting setup,
e.g., three-point softbox setup] to [lighting purpose]. The camera angle is
a [angle type] to showcase [specific feature]. Ultra-realistic, with sharp
focus on [key detail]. [Aspect ratio].
Prompt
A high-resolution, studio-lit product photograph of a minimalist ceramic
coffee mug in matte black, presented on a polished concrete surface. The
lighting is a three-point softbox setup designed to create soft, diffused
highlights and eliminate harsh shadows. The camera angle is a slightly
elevated 45-degree shot to showcase its clean lines. Ultra-realistic, with
sharp focus on the steam rising from the coffee. Square image.
Python
from google import genai
from google.genai import types
client = genai.Client()
response = client.models.generate_content(
model="gemini-2.5-flash-image",
contents="A high-resolution, studio-lit product photograph of a minimalist ceramic coffee mug in matte black, presented on a polished concrete surface. The lighting is a three-point softbox setup designed to create soft, diffused highlights and eliminate harsh shadows. The camera angle is a slightly elevated 45-degree shot to showcase its clean lines. Ultra-realistic, with sharp focus on the steam rising from the coffee. Square image.",
)
for part in response.parts:
if part.text is not None:
print(part.text)
elif part.inline_data is not None:
image = part.as_image()
image.save("product_mockup.png")
JavaScript
import { GoogleGenAI } from "@google/genai";
import * as fs from "node:fs";
async function main() {
const ai = new GoogleGenAI({});
const prompt =
"A high-resolution, studio-lit product photograph of a minimalist ceramic coffee mug in matte black, presented on a polished concrete surface. The lighting is a three-point softbox setup designed to create soft, diffused highlights and eliminate harsh shadows. The camera angle is a slightly elevated 45-degree shot to showcase its clean lines. Ultra-realistic, with sharp focus on the steam rising from the coffee. Square image.";
const response = await ai.models.generateContent({
model: "gemini-2.5-flash-image",
contents: prompt,
});
for (const part of response.candidates[0].content.parts) {
if (part.text) {
console.log(part.text);
} else if (part.inlineData) {
const imageData = part.inlineData.data;
const buffer = Buffer.from(imageData, "base64");
fs.writeFileSync("product_mockup.png", buffer);
console.log("Image saved as product_mockup.png");
}
}
}
main();
Go
package main
import (
"context"
"fmt"
"log"
"os"
"google.golang.org/genai"
)
func main() {
ctx := context.Background()
client, err := genai.NewClient(ctx, nil)
if err != nil {
log.Fatal(err)
}
result, _ := client.Models.GenerateContent(
ctx,
"gemini-2.5-flash-image",
genai.Text("A high-resolution, studio-lit product photograph of a minimalist ceramic coffee mug in matte black, presented on a polished concrete surface. The lighting is a three-point softbox setup designed to create soft, diffused highlights and eliminate harsh shadows. The camera angle is a slightly elevated 45-degree shot to showcase its clean lines. Ultra-realistic, with sharp focus on the steam rising from the coffee. Square image."),
)
for _, part := range result.Candidates[0].Content.Parts {
if part.Text != "" {
fmt.Println(part.Text)
} else if part.InlineData != nil {
imageBytes := part.InlineData.Data
outputFilename := "product_mockup.png"
_ = os.WriteFile(outputFilename, imageBytes, 0644)
}
}
}
REST
curl -s -X POST
"https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-flash-image:generateContent" \
-H "x-goog-api-key: $GEMINI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"contents": [{
"parts": [
{"text": "A high-resolution, studio-lit product photograph of a minimalist ceramic coffee mug in matte black, presented on a polished concrete surface. The lighting is a three-point softbox setup designed to create soft, diffused highlights and eliminate harsh shadows. The camera angle is a slightly elevated 45-degree shot to showcase its clean lines. Ultra-realistic, with sharp focus on the steam rising from the coffee. Square image."}
]
}]
}' \
| grep -o '"data": "[^"]*"' \
| cut -d'"' -f4 \
| base64 --decode > product_mockup.png
5. Design minimaliste et espace négatif
Idéal pour créer des arrière-plans pour des sites Web, des présentations ou des supports marketing sur lesquels du texte sera superposé.
Modèle
A minimalist composition featuring a single [subject] positioned in the
[bottom-right/top-left/etc.] of the frame. The background is a vast, empty
[color] canvas, creating significant negative space. Soft, subtle lighting.
[Aspect ratio].
Prompt
A minimalist composition featuring a single, delicate red maple leaf
positioned in the bottom-right of the frame. The background is a vast, empty
off-white canvas, creating significant negative space for text. Soft,
diffused lighting from the top left. Square image.
Python
from google import genai
from google.genai import types
client = genai.Client()
response = client.models.generate_content(
model="gemini-2.5-flash-image",
contents="A minimalist composition featuring a single, delicate red maple leaf positioned in the bottom-right of the frame. The background is a vast, empty off-white canvas, creating significant negative space for text. Soft, diffused lighting from the top left. Square image.",
)
for part in response.parts:
if part.text is not None:
print(part.text)
elif part.inline_data is not None:
image = part.as_image()
image.save("minimalist_design.png")
JavaScript
import { GoogleGenAI } from "@google/genai";
import * as fs from "node:fs";
async function main() {
const ai = new GoogleGenAI({});
const prompt =
"A minimalist composition featuring a single, delicate red maple leaf positioned in the bottom-right of the frame. The background is a vast, empty off-white canvas, creating significant negative space for text. Soft, diffused lighting from the top left. Square image.";
const response = await ai.models.generateContent({
model: "gemini-2.5-flash-image",
contents: prompt,
});
for (const part of response.candidates[0].content.parts) {
if (part.text) {
console.log(part.text);
} else if (part.inlineData) {
const imageData = part.inlineData.data;
const buffer = Buffer.from(imageData, "base64");
fs.writeFileSync("minimalist_design.png", buffer);
console.log("Image saved as minimalist_design.png");
}
}
}
main();
Go
package main
import (
"context"
"fmt"
"log"
"os"
"google.golang.org/genai"
)
func main() {
ctx := context.Background()
client, err := genai.NewClient(ctx, nil)
if err != nil {
log.Fatal(err)
}
result, _ := client.Models.GenerateContent(
ctx,
"gemini-2.5-flash-image",
genai.Text("A minimalist composition featuring a single, delicate red maple leaf positioned in the bottom-right of the frame. The background is a vast, empty off-white canvas, creating significant negative space for text. Soft, diffused lighting from the top left. Square image."),
)
for _, part := range result.Candidates[0].Content.Parts {
if part.Text != "" {
fmt.Println(part.Text)
} else if part.InlineData != nil {
imageBytes := part.InlineData.Data
outputFilename := "minimalist_design.png"
_ = os.WriteFile(outputFilename, imageBytes, 0644)
}
}
}
REST
curl -s -X POST
"https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-flash-image:generateContent" \
-H "x-goog-api-key: $GEMINI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"contents": [{
"parts": [
{"text": "A minimalist composition featuring a single, delicate red maple leaf positioned in the bottom-right of the frame. The background is a vast, empty off-white canvas, creating significant negative space for text. Soft, diffused lighting from the top left. Square image."}
]
}]
}' \
| grep -o '"data": "[^"]*"' \
| cut -d'"' -f4 \
| base64 --decode > minimalist_design.png
6. Art séquentiel (panneau de bande dessinée / storyboard)
S'appuie sur la cohérence des personnages et la description des scènes pour créer des panneaux de narration visuelle.
Modèle
A single comic book panel in a [art style] style. In the foreground,
[character description and action]. In the background, [setting details].
The panel has a [dialogue/caption box] with the text "[Text]". The lighting
creates a [mood] mood. [Aspect ratio].
Prompt
A single comic book panel in a gritty, noir art style with high-contrast
black and white inks. In the foreground, a detective in a trench coat stands
under a flickering streetlamp, rain soaking his shoulders. In the
background, the neon sign of a desolate bar reflects in a puddle. A caption
box at the top reads "The city was a tough place to keep secrets." The
lighting is harsh, creating a dramatic, somber mood. Landscape.
Python
from google import genai
from google.genai import types
client = genai.Client()
response = client.models.generate_content(
model="gemini-2.5-flash-image",
contents="A single comic book panel in a gritty, noir art style with high-contrast black and white inks. In the foreground, a detective in a trench coat stands under a flickering streetlamp, rain soaking his shoulders. In the background, the neon sign of a desolate bar reflects in a puddle. A caption box at the top reads \"The city was a tough place to keep secrets.\" The lighting is harsh, creating a dramatic, somber mood. Landscape.",
)
for part in response.parts:
if part.text is not None:
print(part.text)
elif part.inline_data is not None:
image = part.as_image()
image.save("comic_panel.png")
JavaScript
import { GoogleGenAI } from "@google/genai";
import * as fs from "node:fs";
async function main() {
const ai = new GoogleGenAI({});
const prompt =
"A single comic book panel in a gritty, noir art style with high-contrast black and white inks. In the foreground, a detective in a trench coat stands under a flickering streetlamp, rain soaking his shoulders. In the background, the neon sign of a desolate bar reflects in a puddle. A caption box at the top reads \"The city was a tough place to keep secrets.\" The lighting is harsh, creating a dramatic, somber mood. Landscape.";
const response = await ai.models.generateContent({
model: "gemini-2.5-flash-image",
contents: prompt,
});
for (const part of response.candidates[0].content.parts) {
if (part.text) {
console.log(part.text);
} else if (part.inlineData) {
const imageData = part.inlineData.data;
const buffer = Buffer.from(imageData, "base64");
fs.writeFileSync("comic_panel.png", buffer);
console.log("Image saved as comic_panel.png");
}
}
}
main();
Go
package main
import (
"context"
"fmt"
"log"
"os"
"google.golang.org/genai"
)
func main() {
ctx := context.Background()
client, err := genai.NewClient(ctx, nil)
if err != nil {
log.Fatal(err)
}
result, _ := client.Models.GenerateContent(
ctx,
"gemini-2.5-flash-image",
genai.Text("A single comic book panel in a gritty, noir art style with high-contrast black and white inks. In the foreground, a detective in a trench coat stands under a flickering streetlamp, rain soaking his shoulders. In the background, the neon sign of a desolate bar reflects in a puddle. A caption box at the top reads \"The city was a tough place to keep secrets.\" The lighting is harsh, creating a dramatic, somber mood. Landscape."),
)
for _, part := range result.Candidates[0].Content.Parts {
if part.Text != "" {
fmt.Println(part.Text)
} else if part.InlineData != nil {
imageBytes := part.InlineData.Data
outputFilename := "comic_panel.png"
_ = os.WriteFile(outputFilename, imageBytes, 0644)
}
}
}
REST
curl -s -X POST
"https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-flash-image:generateContent" \
-H "x-goog-api-key: $GEMINI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"contents": [{
"parts": [
{"text": "A single comic book panel in a gritty, noir art style with high-contrast black and white inks. In the foreground, a detective in a trench coat stands under a flickering streetlamp, rain soaking his shoulders. In the background, the neon sign of a desolate bar reflects in a puddle. A caption box at the top reads \"The city was a tough place to keep secrets.\" The lighting is harsh, creating a dramatic, somber mood. Landscape."}
]
}]
}' \
| grep -o '"data": "[^"]*"' \
| cut -d'"' -f4 \
| base64 --decode > comic_panel.png
Requêtes pour modifier des images
Ces exemples montrent comment fournir des images en plus de vos requêtes textuelles pour l'édition, la composition et le transfert de style.
1. Ajouter et supprimer des éléments
Fournissez une image et décrivez la modification que vous souhaitez apporter. Le modèle correspondra au style, à la luminosité et à la perspective de l'image d'origine.
Modèle
Using the provided image of [subject], please [add/remove/modify] [element]
to/from the scene. Ensure the change is [description of how the change should
integrate].
Prompt
"Using the provided image of my cat, please add a small, knitted wizard hat
on its head. Make it look like it's sitting comfortably and matches the soft
lighting of the photo."
Python
from google import genai
from google.genai import types
from PIL import Image
client = genai.Client()
# Base image prompt: "A photorealistic picture of a fluffy ginger cat sitting on a wooden floor, looking directly at the camera. Soft, natural light from a window."
image_input = Image.open('/path/to/your/cat_photo.png')
text_input = """Using the provided image of my cat, please add a small, knitted wizard hat on its head. Make it look like it's sitting comfortably and not falling off."""
# Generate an image from a text prompt
response = client.models.generate_content(
model="gemini-2.5-flash-image",
contents=[text_input, image_input],
)
for part in response.parts:
if part.text is not None:
print(part.text)
elif part.inline_data is not None:
image = part.as_image()
image.save("cat_with_hat.png")
JavaScript
import { GoogleGenAI } from "@google/genai";
import * as fs from "node:fs";
async function main() {
const ai = new GoogleGenAI({});
const imagePath = "/path/to/your/cat_photo.png";
const imageData = fs.readFileSync(imagePath);
const base64Image = imageData.toString("base64");
const prompt = [
{ text: "Using the provided image of my cat, please add a small, knitted wizard hat on its head. Make it look like it's sitting comfortably and not falling off." },
{
inlineData: {
mimeType: "image/png",
data: base64Image,
},
},
];
const response = await ai.models.generateContent({
model: "gemini-2.5-flash-image",
contents: prompt,
});
for (const part of response.candidates[0].content.parts) {
if (part.text) {
console.log(part.text);
} else if (part.inlineData) {
const imageData = part.inlineData.data;
const buffer = Buffer.from(imageData, "base64");
fs.writeFileSync("cat_with_hat.png", buffer);
console.log("Image saved as cat_with_hat.png");
}
}
}
main();
Go
package main
import (
"context"
"fmt"
"log"
"os"
"google.golang.org/genai"
)
func main() {
ctx := context.Background()
client, err := genai.NewClient(ctx, nil)
if err != nil {
log.Fatal(err)
}
imagePath := "/path/to/your/cat_photo.png"
imgData, _ := os.ReadFile(imagePath)
parts := []*genai.Part{
genai.NewPartFromText("Using the provided image of my cat, please add a small, knitted wizard hat on its head. Make it look like it's sitting comfortably and not falling off."),
&genai.Part{
InlineData: &genai.Blob{
MIMEType: "image/png",
Data: imgData,
},
},
}
contents := []*genai.Content{
genai.NewContentFromParts(parts, genai.RoleUser),
}
result, _ := client.Models.GenerateContent(
ctx,
"gemini-2.5-flash-image",
contents,
)
for _, part := range result.Candidates[0].Content.Parts {
if part.Text != "" {
fmt.Println(part.Text)
} else if part.InlineData != nil {
imageBytes := part.InlineData.Data
outputFilename := "cat_with_hat.png"
_ = os.WriteFile(outputFilename, imageBytes, 0644)
}
}
}
REST
IMG_PATH=/path/to/your/cat_photo.png
if [[ "$(base64 --version 2>&1)" = *"FreeBSD"* ]]; then
B64FLAGS="--input"
else
B64FLAGS="-w0"
fi
IMG_BASE64=$(base64 "$B64FLAGS" "$IMG_PATH" 2>&1)
curl -X POST \
"https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-flash-image:generateContent" \
-H "x-goog-api-key: $GEMINI_API_KEY" \
-H 'Content-Type: application/json' \
-d "{
\"contents\": [{
\"parts\":[
{\"text\": \"Using the provided image of my cat, please add a small, knitted wizard hat on its head. Make it look like it's sitting comfortably and not falling off.\"},
{
\"inline_data\": {
\"mime_type\":\"image/png\",
\"data\": \"$IMG_BASE64\"
}
}
]
}]
}" \
| grep -o '"data": "[^"]*"' \
| cut -d'"' -f4 \
| base64 --decode > cat_with_hat.png
Entrée |
Sortie |
|
|
2. Remplissage (masquage sémantique)
Définissez de manière conversationnelle un "masque" pour modifier une partie spécifique d'une image tout en laissant le reste intact.
Modèle
Using the provided image, change only the [specific element] to [new
element/description]. Keep everything else in the image exactly the same,
preserving the original style, lighting, and composition.
Prompt
"Using the provided image of a living room, change only the blue sofa to be
a vintage, brown leather chesterfield sofa. Keep the rest of the room,
including the pillows on the sofa and the lighting, unchanged."
Python
from google import genai
from google.genai import types
from PIL import Image
client = genai.Client()
# Base image prompt: "A wide shot of a modern, well-lit living room with a prominent blue sofa in the center. A coffee table is in front of it and a large window is in the background."
living_room_image = Image.open('/path/to/your/living_room.png')
text_input = """Using the provided image of a living room, change only the blue sofa to be a vintage, brown leather chesterfield sofa. Keep the rest of the room, including the pillows on the sofa and the lighting, unchanged."""
# Generate an image from a text prompt
response = client.models.generate_content(
model="gemini-2.5-flash-image",
contents=[living_room_image, text_input],
)
for part in response.parts:
if part.text is not None:
print(part.text)
elif part.inline_data is not None:
image = part.as_image()
image.save("living_room_edited.png")
JavaScript
import { GoogleGenAI } from "@google/genai";
import * as fs from "node:fs";
async function main() {
const ai = new GoogleGenAI({});
const imagePath = "/path/to/your/living_room.png";
const imageData = fs.readFileSync(imagePath);
const base64Image = imageData.toString("base64");
const prompt = [
{
inlineData: {
mimeType: "image/png",
data: base64Image,
},
},
{ text: "Using the provided image of a living room, change only the blue sofa to be a vintage, brown leather chesterfield sofa. Keep the rest of the room, including the pillows on the sofa and the lighting, unchanged." },
];
const response = await ai.models.generateContent({
model: "gemini-2.5-flash-image",
contents: prompt,
});
for (const part of response.candidates[0].content.parts) {
if (part.text) {
console.log(part.text);
} else if (part.inlineData) {
const imageData = part.inlineData.data;
const buffer = Buffer.from(imageData, "base64");
fs.writeFileSync("living_room_edited.png", buffer);
console.log("Image saved as living_room_edited.png");
}
}
}
main();
Go
package main
import (
"context"
"fmt"
"log"
"os"
"google.golang.org/genai"
)
func main() {
ctx := context.Background()
client, err := genai.NewClient(ctx, nil)
if err != nil {
log.Fatal(err)
}
imagePath := "/path/to/your/living_room.png"
imgData, _ := os.ReadFile(imagePath)
parts := []*genai.Part{
&genai.Part{
InlineData: &genai.Blob{
MIMEType: "image/png",
Data: imgData,
},
},
genai.NewPartFromText("Using the provided image of a living room, change only the blue sofa to be a vintage, brown leather chesterfield sofa. Keep the rest of the room, including the pillows on the sofa and the lighting, unchanged."),
}
contents := []*genai.Content{
genai.NewContentFromParts(parts, genai.RoleUser),
}
result, _ := client.Models.GenerateContent(
ctx,
"gemini-2.5-flash-image",
contents,
)
for _, part := range result.Candidates[0].Content.Parts {
if part.Text != "" {
fmt.Println(part.Text)
} else if part.InlineData != nil {
imageBytes := part.InlineData.Data
outputFilename := "living_room_edited.png"
_ = os.WriteFile(outputFilename, imageBytes, 0644)
}
}
}
REST
IMG_PATH=/path/to/your/living_room.png
if [[ "$(base64 --version 2>&1)" = *"FreeBSD"* ]]; then
B64FLAGS="--input"
else
B64FLAGS="-w0"
fi
IMG_BASE64=$(base64 "$B64FLAGS" "$IMG_PATH" 2>&1)
curl -X POST \
"https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-flash-image:generateContent" \
-H "x-goog-api-key: $GEMINI_API_KEY" \
-H 'Content-Type: application/json' \
-d "{
\"contents\": [{
\"parts\":[
{
\"inline_data\": {
\"mime_type\":\"image/png\",
\"data\": \"$IMG_BASE64\"
}
},
{\"text\": \"Using the provided image of a living room, change only the blue sofa to be a vintage, brown leather chesterfield sofa. Keep the rest of the room, including the pillows on the sofa and the lighting, unchanged.\"}
]
}]
}" \
| grep -o '"data": "[^"]*"' \
| cut -d'"' -f4 \
| base64 --decode > living_room_edited.png
Entrée |
Sortie |
|
|
3. Transfert de style
Fournissez une image et demandez au modèle de recréer son contenu dans un autre style artistique.
Modèle
Transform the provided photograph of [subject] into the artistic style of [artist/art style]. Preserve the original composition but render it with [description of stylistic elements].
Prompt
"Transform the provided photograph of a modern city street at night into the artistic style of Vincent van Gogh's 'Starry Night'. Preserve the original composition of buildings and cars, but render all elements with swirling, impasto brushstrokes and a dramatic palette of deep blues and bright yellows."
Python
from google import genai
from google.genai import types
from PIL import Image
client = genai.Client()
# Base image prompt: "A photorealistic, high-resolution photograph of a busy city street in New York at night, with bright neon signs, yellow taxis, and tall skyscrapers."
city_image = Image.open('/path/to/your/city.png')
text_input = """Transform the provided photograph of a modern city street at night into the artistic style of Vincent van Gogh's 'Starry Night'. Preserve the original composition of buildings and cars, but render all elements with swirling, impasto brushstrokes and a dramatic palette of deep blues and bright yellows."""
# Generate an image from a text prompt
response = client.models.generate_content(
model="gemini-2.5-flash-image",
contents=[city_image, text_input],
)
for part in response.parts:
if part.text is not None:
print(part.text)
elif part.inline_data is not None:
image = part.as_image()
image.save("city_style_transfer.png")
JavaScript
import { GoogleGenAI } from "@google/genai";
import * as fs from "node:fs";
async function main() {
const ai = new GoogleGenAI({});
const imagePath = "/path/to/your/city.png";
const imageData = fs.readFileSync(imagePath);
const base64Image = imageData.toString("base64");
const prompt = [
{
inlineData: {
mimeType: "image/png",
data: base64Image,
},
},
{ text: "Transform the provided photograph of a modern city street at night into the artistic style of Vincent van Gogh's 'Starry Night'. Preserve the original composition of buildings and cars, but render all elements with swirling, impasto brushstrokes and a dramatic palette of deep blues and bright yellows." },
];
const response = await ai.models.generateContent({
model: "gemini-2.5-flash-image",
contents: prompt,
});
for (const part of response.candidates[0].content.parts) {
if (part.text) {
console.log(part.text);
} else if (part.inlineData) {
const imageData = part.inlineData.data;
const buffer = Buffer.from(imageData, "base64");
fs.writeFileSync("city_style_transfer.png", buffer);
console.log("Image saved as city_style_transfer.png");
}
}
}
main();
Go
package main
import (
"context"
"fmt"
"log"
"os"
"google.golang.org/genai"
)
func main() {
ctx := context.Background()
client, err := genai.NewClient(ctx, nil)
if err != nil {
log.Fatal(err)
}
imagePath := "/path/to/your/city.png"
imgData, _ := os.ReadFile(imagePath)
parts := []*genai.Part{
&genai.Part{
InlineData: &genai.Blob{
MIMEType: "image/png",
Data: imgData,
},
},
genai.NewPartFromText("Transform the provided photograph of a modern city street at night into the artistic style of Vincent van Gogh's 'Starry Night'. Preserve the original composition of buildings and cars, but render all elements with swirling, impasto brushstrokes and a dramatic palette of deep blues and bright yellows."),
}
contents := []*genai.Content{
genai.NewContentFromParts(parts, genai.RoleUser),
}
result, _ := client.Models.GenerateContent(
ctx,
"gemini-2.5-flash-image",
contents,
)
for _, part := range result.Candidates[0].Content.Parts {
if part.Text != "" {
fmt.Println(part.Text)
} else if part.InlineData != nil {
imageBytes := part.InlineData.Data
outputFilename := "city_style_transfer.png"
_ = os.WriteFile(outputFilename, imageBytes, 0644)
}
}
}
REST
IMG_PATH=/path/to/your/city.png
if [[ "$(base64 --version 2>&1)" = *"FreeBSD"* ]]; then
B64FLAGS="--input"
else
B64FLAGS="-w0"
fi
IMG_BASE64=$(base64 "$B64FLAGS" "$IMG_PATH" 2>&1)
curl -X POST \
"https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-flash-image:generateContent" \
-H "x-goog-api-key: $GEMINI_API_KEY" \
-H 'Content-Type: application/json' \
-d "{
\"contents\": [{
\"parts\":[
{
\"inline_data\": {
\"mime_type\":\"image/png\",
\"data\": \"$IMG_BASE64\"
}
},
{\"text\": \"Transform the provided photograph of a modern city street at night into the artistic style of Vincent van Gogh's 'Starry Night'. Preserve the original composition of buildings and cars, but render all elements with swirling, impasto brushstrokes and a dramatic palette of deep blues and bright yellows.\"}
]
}]
}" \
| grep -o '"data": "[^"]*"' \
| cut -d'"' -f4 \
| base64 --decode > city_style_transfer.png
Entrée |
Sortie |
|
|
4. Composition avancée : combiner plusieurs images
Fournissez plusieurs images comme contexte pour créer une nouvelle scène composite. C'est l'idéal pour les maquettes de produits ou les collages créatifs.
Modèle
Create a new image by combining the elements from the provided images. Take
the [element from image 1] and place it with/on the [element from image 2].
The final image should be a [description of the final scene].
Prompt
"Create a professional e-commerce fashion photo. Take the blue floral dress
from the first image and let the woman from the second image wear it.
Generate a realistic, full-body shot of the woman wearing the dress, with
the lighting and shadows adjusted to match the outdoor environment."
Python
from google import genai
from google.genai import types
from PIL import Image
client = genai.Client()
# Base image prompts:
# 1. Dress: "A professionally shot photo of a blue floral summer dress on a plain white background, ghost mannequin style."
# 2. Model: "Full-body shot of a woman with her hair in a bun, smiling, standing against a neutral grey studio background."
dress_image = Image.open('/path/to/your/dress.png')
model_image = Image.open('/path/to/your/model.png')
text_input = """Create a professional e-commerce fashion photo. Take the blue floral dress from the first image and let the woman from the second image wear it. Generate a realistic, full-body shot of the woman wearing the dress, with the lighting and shadows adjusted to match the outdoor environment."""
# Generate an image from a text prompt
response = client.models.generate_content(
model="gemini-2.5-flash-image",
contents=[dress_image, model_image, text_input],
)
for part in response.parts:
if part.text is not None:
print(part.text)
elif part.inline_data is not None:
image = part.as_image()
image.save("fashion_ecommerce_shot.png")
JavaScript
import { GoogleGenAI } from "@google/genai";
import * as fs from "node:fs";
async function main() {
const ai = new GoogleGenAI({});
const imagePath1 = "/path/to/your/dress.png";
const imageData1 = fs.readFileSync(imagePath1);
const base64Image1 = imageData1.toString("base64");
const imagePath2 = "/path/to/your/model.png";
const imageData2 = fs.readFileSync(imagePath2);
const base64Image2 = imageData2.toString("base64");
const prompt = [
{
inlineData: {
mimeType: "image/png",
data: base64Image1,
},
},
{
inlineData: {
mimeType: "image/png",
data: base64Image2,
},
},
{ text: "Create a professional e-commerce fashion photo. Take the blue floral dress from the first image and let the woman from the second image wear it. Generate a realistic, full-body shot of the woman wearing the dress, with the lighting and shadows adjusted to match the outdoor environment." },
];
const response = await ai.models.generateContent({
model: "gemini-2.5-flash-image",
contents: prompt,
});
for (const part of response.candidates[0].content.parts) {
if (part.text) {
console.log(part.text);
} else if (part.inlineData) {
const imageData = part.inlineData.data;
const buffer = Buffer.from(imageData, "base64");
fs.writeFileSync("fashion_ecommerce_shot.png", buffer);
console.log("Image saved as fashion_ecommerce_shot.png");
}
}
}
main();
Go
package main
import (
"context"
"fmt"
"log"
"os"
"google.golang.org/genai"
)
func main() {
ctx := context.Background()
client, err := genai.NewClient(ctx, nil)
if err != nil {
log.Fatal(err)
}
imgData1, _ := os.ReadFile("/path/to/your/dress.png")
imgData2, _ := os.ReadFile("/path/to/your/model.png")
parts := []*genai.Part{
&genai.Part{
InlineData: &genai.Blob{
MIMEType: "image/png",
Data: imgData1,
},
},
&genai.Part{
InlineData: &genai.Blob{
MIMEType: "image/png",
Data: imgData2,
},
},
genai.NewPartFromText("Create a professional e-commerce fashion photo. Take the blue floral dress from the first image and let the woman from the second image wear it. Generate a realistic, full-body shot of the woman wearing the dress, with the lighting and shadows adjusted to match the outdoor environment."),
}
contents := []*genai.Content{
genai.NewContentFromParts(parts, genai.RoleUser),
}
result, _ := client.Models.GenerateContent(
ctx,
"gemini-2.5-flash-image",
contents,
)
for _, part := range result.Candidates[0].Content.Parts {
if part.Text != "" {
fmt.Println(part.Text)
} else if part.InlineData != nil {
imageBytes := part.InlineData.Data
outputFilename := "fashion_ecommerce_shot.png"
_ = os.WriteFile(outputFilename, imageBytes, 0644)
}
}
}
REST
IMG_PATH1=/path/to/your/dress.png
IMG_PATH2=/path/to/your/model.png
if [[ "$(base64 --version 2>&1)" = *"FreeBSD"* ]]; then
B64FLAGS="--input"
else
B64FLAGS="-w0"
fi
IMG1_BASE64=$(base64 "$B64FLAGS" "$IMG_PATH1" 2>&1)
IMG2_BASE64=$(base64 "$B64FLAGS" "$IMG_PATH2" 2>&1)
curl -X POST \
"https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-flash-image:generateContent" \
-H "x-goog-api-key: $GEMINI_API_KEY" \
-H 'Content-Type: application/json' \
-d "{
\"contents\": [{
\"parts\":[
{
\"inline_data\": {
\"mime_type\":\"image/png\",
\"data\": \"$IMG1_BASE64\"
}
},
{
\"inline_data\": {
\"mime_type\":\"image/png\",
\"data\": \"$IMG2_BASE64\"
}
},
{\"text\": \"Create a professional e-commerce fashion photo. Take the blue floral dress from the first image and let the woman from the second image wear it. Generate a realistic, full-body shot of the woman wearing the dress, with the lighting and shadows adjusted to match the outdoor environment.\"}
]
}]
}" \
| grep -o '"data": "[^"]*"' \
| cut -d'"' -f4 \
| base64 --decode > fashion_ecommerce_shot.png
Saisie 1 |
Saisie 2 |
Sortie |
|
|
|
5. Préservation des détails haute fidélité
Pour vous assurer que les détails importants (comme un visage ou un logo) sont conservés lors d'une modification, décrivez-les très précisément dans votre demande de modification.
Modèle
Using the provided images, place [element from image 2] onto [element from
image 1]. Ensure that the features of [element from image 1] remain
completely unchanged. The added element should [description of how the
element should integrate].
Prompt
"Take the first image of the woman with brown hair, blue eyes, and a neutral
expression. Add the logo from the second image onto her black t-shirt.
Ensure the woman's face and features remain completely unchanged. The logo
should look like it's naturally printed on the fabric, following the folds
of the shirt."
Python
from google import genai
from google.genai import types
from PIL import Image
client = genai.Client()
# Base image prompts:
# 1. Woman: "A professional headshot of a woman with brown hair and blue eyes, wearing a plain black t-shirt, against a neutral studio background."
# 2. Logo: "A simple, modern logo with the letters 'G' and 'A' in a white circle."
woman_image = Image.open('/path/to/your/woman.png')
logo_image = Image.open('/path/to/your/logo.png')
text_input = """Take the first image of the woman with brown hair, blue eyes, and a neutral expression. Add the logo from the second image onto her black t-shirt. Ensure the woman's face and features remain completely unchanged. The logo should look like it's naturally printed on the fabric, following the folds of the shirt."""
# Generate an image from a text prompt
response = client.models.generate_content(
model="gemini-2.5-flash-image",
contents=[woman_image, logo_image, text_input],
)
for part in response.parts:
if part.text is not None:
print(part.text)
elif part.inline_data is not None:
image = part.as_image()
image.save("woman_with_logo.png")
JavaScript
import { GoogleGenAI } from "@google/genai";
import * as fs from "node:fs";
async function main() {
const ai = new GoogleGenAI({});
const imagePath1 = "/path/to/your/woman.png";
const imageData1 = fs.readFileSync(imagePath1);
const base64Image1 = imageData1.toString("base64");
const imagePath2 = "/path/to/your/logo.png";
const imageData2 = fs.readFileSync(imagePath2);
const base64Image2 = imageData2.toString("base64");
const prompt = [
{
inlineData: {
mimeType: "image/png",
data: base64Image1,
},
},
{
inlineData: {
mimeType: "image/png",
data: base64Image2,
},
},
{ text: "Take the first image of the woman with brown hair, blue eyes, and a neutral expression. Add the logo from the second image onto her black t-shirt. Ensure the woman's face and features remain completely unchanged. The logo should look like it's naturally printed on the fabric, following the folds of the shirt." },
];
const response = await ai.models.generateContent({
model: "gemini-2.5-flash-image",
contents: prompt,
});
for (const part of response.candidates[0].content.parts) {
if (part.text) {
console.log(part.text);
} else if (part.inlineData) {
const imageData = part.inlineData.data;
const buffer = Buffer.from(imageData, "base64");
fs.writeFileSync("woman_with_logo.png", buffer);
console.log("Image saved as woman_with_logo.png");
}
}
}
main();
Go
package main
import (
"context"
"fmt"
"log"
"os"
"google.golang.org/genai"
)
func main() {
ctx := context.Background()
client, err := genai.NewClient(ctx, nil)
if err != nil {
log.Fatal(err)
}
imgData1, _ := os.ReadFile("/path/to/your/woman.png")
imgData2, _ := os.ReadFile("/path/to/your/logo.png")
parts := []*genai.Part{
&genai.Part{
InlineData: &genai.Blob{
MIMEType: "image/png",
Data: imgData1,
},
},
&genai.Part{
InlineData: &genai.Blob{
MIMEType: "image/png",
Data: imgData2,
},
},
genai.NewPartFromText("Take the first image of the woman with brown hair, blue eyes, and a neutral expression. Add the logo from the second image onto her black t-shirt. Ensure the woman's face and features remain completely unchanged. The logo should look like it's naturally printed on the fabric, following the folds of the shirt."),
}
contents := []*genai.Content{
genai.NewContentFromParts(parts, genai.RoleUser),
}
result, _ := client.Models.GenerateContent(
ctx,
"gemini-2.5-flash-image",
contents,
)
for _, part := range result.Candidates[0].Content.Parts {
if part.Text != "" {
fmt.Println(part.Text)
} else if part.InlineData != nil {
imageBytes := part.InlineData.Data
outputFilename := "woman_with_logo.png"
_ = os.WriteFile(outputFilename, imageBytes, 0644)
}
}
}
REST
IMG_PATH1=/path/to/your/woman.png
IMG_PATH2=/path/to/your/logo.png
if [[ "$(base64 --version 2>&1)" = *"FreeBSD"* ]]; then
B64FLAGS="--input"
else
B64FLAGS="-w0"
fi
IMG1_BASE64=$(base64 "$B64FLAGS" "$IMG_PATH1" 2>&1)
IMG2_BASE64=$(base64 "$B64FLAGS" "$IMG_PATH2" 2>&1)
curl -X POST \
"https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-flash-image:generateContent" \
-H "x-goog-api-key: $GEMINI_API_KEY" \
-H 'Content-Type: application/json' \
-d "{
\"contents\": [{
\"parts\":[
{
\"inline_data\": {
\"mime_type\":\"image/png\",
\"data\": \"$IMG1_BASE64\"
}
},
{
\"inline_data\": {
\"mime_type\":\"image/png\",
\"data\": \"$IMG2_BASE64\"
}
},
{\"text\": \"Take the first image of the woman with brown hair, blue eyes, and a neutral expression. Add the logo from the second image onto her black t-shirt. Ensure the woman's face and features remain completely unchanged. The logo should look like it's naturally printed on the fabric, following the folds of the shirt.\"}
]
}]
}" \
| grep -o '"data": "[^"]*"' \
| cut -d'"' -f4 \
| base64 --decode > woman_with_logo.png
Saisie 1 |
Saisie 2 |
Sortie |
|
|
|
Bonnes pratiques
Pour passer de bons résultats à d'excellents résultats, intégrez ces stratégies professionnelles à votre workflow.
- Soyez très précis : plus vous fournissez de détails, plus vous avez de contrôle. Au lieu de "armure fantastique", décrivez-la : "armure de plates elfique ornée, gravée de motifs en feuille d'argent, avec un col montant et des épaulières en forme d'ailes de faucon".
- Fournissez le contexte et l'intention : expliquez la finalité de l'image. La compréhension du contexte par le modèle influencera le résultat final. Par exemple, "Crée un logo pour une marque de soins pour la peau haut de gamme et minimaliste" donnera de meilleurs résultats que "Crée un logo".
- Répétez et affinez : ne vous attendez pas à obtenir une image parfaite du premier coup. Utilisez la nature conversationnelle du modèle pour apporter de petites modifications. Faites un suivi avec des requêtes telles que "C'est super, mais peux-tu rendre l'éclairage un peu plus chaud ?" ou "Garde tout pareil, mais rends l'expression du personnage plus sérieuse".
- Utilisez des instructions détaillées : pour les scènes complexes comportant de nombreux éléments, divisez votre requête en étapes. "Commence par créer un arrière-plan représentant une forêt sereine et brumeuse à l'aube. Ensuite, au premier plan, ajoutez un ancien autel en pierre recouvert de mousse. Enfin, placez une épée lumineuse sur l'autel."
- Utilisez des requêtes négatives sémantiques : au lieu de dire "pas de voitures", décrivez la scène souhaitée de manière positive : "une rue vide et désertée, sans aucun signe de circulation".
- Contrôlez la caméra : utilisez un langage photographique et cinématographique pour contrôler la composition. Termes tels que
wide-angle shot,macro shotetlow-angle perspective.
Limites
- Pour des performances optimales, utilisez les langues suivantes : EN, es-MX, ja-JP, zh-CN, hi-IN.
- La génération d'images n'est pas compatible avec les entrées audio ou vidéo.
- Le modèle ne suit pas toujours le nombre exact d'images de sortie que l'utilisateur demande explicitement.
- Le modèle fonctionne mieux avec un maximum de trois images en entrée.
- Lorsque vous générez du texte pour une image, Gemini fonctionne mieux si vous générez d'abord le texte, puis demandez une image avec le texte.
- Toutes les images générées incluent un filigrane SynthID.
Configurations facultatives
Vous pouvez éventuellement configurer les modalités de réponse et le format de la sortie du modèle dans le champ config des appels generate_content.
Types de sortie
Par défaut, le modèle renvoie des réponses textuelles et imagées (c'est-à-dire response_modalities=['Text', 'Image']). Vous pouvez configurer la réponse pour qu'elle ne renvoie que des images sans texte à l'aide de response_modalities=['Image'].
Python
response = client.models.generate_content(
model="gemini-2.5-flash-image",
contents=[prompt],
config=types.GenerateContentConfig(
response_modalities=['Image']
)
)
JavaScript
const response = await ai.models.generateContent({
model: "gemini-2.5-flash-image",
contents: prompt,
config: {
responseModalities: ['Image']
}
});
Go
result, _ := client.Models.GenerateContent(
ctx,
"gemini-2.5-flash-image",
genai.Text("Create a picture of a nano banana dish in a " +
" fancy restaurant with a Gemini theme"),
&genai.GenerateContentConfig{
ResponseModalities: "Image",
},
)
REST
-d '{
"contents": [{
"parts": [
{"text": "Create a picture of a nano banana dish in a fancy restaurant with a Gemini theme"}
]
}],
"generationConfig": {
"responseModalities": ["Image"]
}
}' \
Formats
Par défaut, le modèle fait correspondre la taille de l'image de sortie à celle de votre image d'entrée, ou génère des carrés au format 1:1.
Vous pouvez contrôler le format de l'image de sortie à l'aide du champ aspect_ratio sous image_config dans la demande de réponse, comme indiqué ici :
Python
response = client.models.generate_content(
model="gemini-2.5-flash-image",
contents=[prompt],
config=types.GenerateContentConfig(
image_config=types.ImageConfig(
aspect_ratio="16:9",
)
)
)
JavaScript
const response = await ai.models.generateContent({
model: "gemini-2.5-flash-image",
contents: prompt,
config: {
imageConfig: {
aspectRatio: "16:9",
},
}
});
Go
result, _ := client.Models.GenerateContent(
ctx,
"gemini-2.5-flash-image",
genai.Text("Create a picture of a nano banana dish in a " +
" fancy restaurant with a Gemini theme"),
&genai.GenerateContentConfig{
ImageConfig: &genai.ImageConfig{
AspectRatio: "16:9",
},
}
)
REST
-d '{
"contents": [{
"parts": [
{"text": "Create a picture of a nano banana dish in a fancy restaurant with a Gemini theme"}
]
}],
"generationConfig": {
"imageConfig": {
"aspectRatio": "16:9"
}
}
}' \
Les différents ratios disponibles et la taille de l'image générée sont listés dans les tableaux suivants :
Gemini 2.5 Flash Image
| Format | Solution | Jetons |
|---|---|---|
| 1:1 | 1024x1024 | 1290 |
| 2:3 | 832 x 1 248 | 1290 |
| 3:2 | 1 248 x 832 | 1290 |
| 3:4 | 864 x 1 184 | 1290 |
| 4:3 | 1 184 x 864 | 1290 |
| 4:5 | 896 x 1 152 | 1290 |
| 5:4 | 1152x896 | 1290 |
| 9:16 | 768 x 1 344 | 1290 |
| 16:9 | 1344x768 | 1290 |
| 21:9 | 1 536 x 672 | 1290 |
Aperçu de l'image Gemini 3 Pro
| Format | Résolution 1K | 1 000 jetons | Résolution 2K | 2 000 jetons | Résolution 4K | Jetons 4K |
|---|---|---|---|---|---|---|
| 1:1 | 1024x1024 | 1210 | 2048 x 2048 | 1210 | 4 096 x 4 096 | 2000 |
| 2:3 | 848 x 1 264 | 1210 | 1696 x 2528 | 1210 | 3392 x 5056 | 2000 |
| 3:2 | 1 264 x 848 | 1210 | 2528 x 1696 | 1210 | 5056 x 3392 | 2000 |
| 3:4 | 896 x 1200 | 1210 | 1792x2400 | 1210 | 3584 x 4800 | 2000 |
| 4:3 | 1 200 x 896 | 1210 | 2400x1792 | 1210 | 4 800 x 3 584 | 2000 |
| 4:5 | 928 x 1 152 | 1210 | 1856 x 2304 | 1210 | 3712 x 4608 | 2000 |
| 5:4 | 1152x928 | 1210 | 2304 x 1856 | 1210 | 4608 x 3712 | 2000 |
| 9:16 | 768 x 1 376 | 1210 | 1536 x 2752 | 1210 | 3072 x 5504 | 2000 |
| 16:9 | 1 376 x 768 | 1210 | 2752 x 1536 | 1210 | 5504 x 3072 | 2000 |
| 21:9 | 1 584 x 672 | 1210 | 3168x1344 | 1210 | 6336 x 2688 | 2000 |
Quand utiliser Imagen ?
En plus d'utiliser les fonctionnalités de génération d'images intégrées à Gemini, vous pouvez également accéder à Imagen, notre modèle spécialisé dans la génération d'images, via l'API Gemini.
| Attribut | Imagen | Image native Gemini |
|---|---|---|
| Points forts | Le modèle est spécialisé dans la génération d'images. | Recommandation par défaut : une flexibilité inégalée, une compréhension contextuelle et une modification simple sans masque. Il est le seul à pouvoir effectuer des modifications conversationnelles multitours. |
| Disponibilité | Disponibilité générale | Aperçu (utilisation en production autorisée) |
| Latence | Faible Optimisé pour des performances en temps quasi réel. | Durée plus longue. Ses fonctionnalités avancées nécessitent davantage de calculs. |
| Coût | Rentable pour les tâches spécialisées. De 0,02 $ par image à 0,12 $par image | Tarification basée sur les jetons. 30 $ par million de jetons pour les images générées (les images sont tokenisées à 1 290 jetons par image, jusqu'à 1 024 x 1 024 px) |
| Tâches recommandées |
|
|
Imagen 4 doit être votre modèle de référence lorsque vous commencez à générer des images avec Imagen. Choisissez Imagen 4 Ultra pour les cas d'utilisation avancés ou lorsque vous avez besoin de la meilleure qualité d'image (notez que vous ne pouvez générer qu'une seule image à la fois).
Étape suivante
- Vous trouverez d'autres exemples et extraits de code dans le guide Cookbook.
- Consultez le guide Veo pour découvrir comment générer des vidéos avec l'API Gemini.
- Pour en savoir plus sur les modèles Gemini, consultez Modèles Gemini.