Référence API
Guides d’intégration
Erreurs
Tarification et facturation
Modèles de texte
Génération de contenu Gemini
Utilisez le format d’API natif Gemini via CometAPI pour la génération de texte, les entrées multimodales, la réflexion/le raisonnement, le function calling, l’ancrage Google Search, le mode JSON et le streaming.
POST
/
v1beta
/
models
/
{model}
:
{operator}
from google import genai
client = genai.Client(
api_key="<COMETAPI_KEY>",
http_options={"api_version": "v1beta", "base_url": "https://api.cometapi.com"},
)
response = client.models.generate_content(
model="gemini-2.5-flash",
contents="Explain how AI works in a few words",
)
print(response.text){
"candidates": [
{
"content": {
"role": "<string>",
"parts": [
{
"text": "<string>",
"functionCall": {
"name": "<string>",
"args": {}
},
"inlineData": {
"mimeType": "<string>",
"data": "<string>"
},
"thought": true
}
]
},
"finishReason": "STOP",
"safetyRatings": [
{
"category": "<string>",
"probability": "<string>",
"blocked": true
}
],
"citationMetadata": {
"citationSources": [
{
"startIndex": 123,
"endIndex": 123,
"uri": "<string>",
"license": "<string>"
}
]
},
"tokenCount": 123,
"avgLogprobs": 123,
"groundingMetadata": {
"groundingChunks": [
{
"web": {
"uri": "<string>",
"title": "<string>"
}
}
],
"groundingSupports": [
{
"groundingChunkIndices": [
123
],
"confidenceScores": [
123
],
"segment": {
"startIndex": 123,
"endIndex": 123,
"text": "<string>"
}
}
],
"webSearchQueries": [
"<string>"
]
},
"index": 123
}
],
"promptFeedback": {
"blockReason": "SAFETY",
"safetyRatings": [
{
"category": "<string>",
"probability": "<string>",
"blocked": true
}
]
},
"usageMetadata": {
"promptTokenCount": 123,
"candidatesTokenCount": 123,
"totalTokenCount": 123,
"trafficType": "<string>",
"thoughtsTokenCount": 123,
"promptTokensDetails": [
{
"modality": "<string>",
"tokenCount": 123
}
],
"candidatesTokensDetails": [
{
"modality": "<string>",
"tokenCount": 123
}
]
},
"modelVersion": "<string>",
"createTime": "<string>",
"responseId": "<string>"
}Vue d’ensemble
CometAPI prend en charge le format d’API natif Gemini, vous donnant un accès complet aux fonctionnalités spécifiques à Gemini comme le contrôle du raisonnement, l’ancrage Google Search, les modalités natives de génération d’images, et plus encore. Utilisez cet endpoint lorsque vous avez besoin de capacités non disponibles via l’endpoint de chat compatible OpenAI.Démarrage rapide
Remplacez l’URL de base et la clé API dans n’importe quel SDK Gemini ou client HTTP :| Paramètre | Valeur par défaut Google | CometAPI |
|---|---|---|
| URL de base | generativelanguage.googleapis.com | api.cometapi.com |
| Clé API | $GEMINI_API_KEY | $COMETAPI_KEY |
Les en-têtes
x-goog-api-key et Authorization: Bearer sont tous deux pris en charge pour l’authentification.Thinking (Reasoning)
Les modèles Gemini peuvent effectuer un raisonnement interne avant de générer une réponse. La méthode de contrôle dépend de la génération du modèle.- Gemini 3 (thinkingLevel)
- Gemini 2.5 (thinkingBudget)
Les modèles Gemini 3 utilisent
thinkingLevel pour contrôler la profondeur du raisonnement. Niveaux disponibles : MINIMAL, LOW, MEDIUM, HIGH.curl "https://api.cometapi.com/v1beta/models/gemini-3.1-pro-preview:generateContent" \
-H "Content-Type: application/json" \
-H "x-goog-api-key: $COMETAPI_KEY" \
-d '{
"contents": [{"parts": [{"text": "Explain quantum physics simply."}]}],
"generationConfig": {
"thinkingConfig": {"thinkingLevel": "LOW"}
}
}'
Les modèles Gemini 2.5 utilisent
thinkingBudget pour un contrôle fin au niveau des token :0— désactiver le raisonnement-1— dynamique (le modèle décide, par défaut)> 0— budget de token spécifique (par exemple,1024,2048)
curl "https://api.cometapi.com/v1beta/models/gemini-2.5-flash:generateContent" \
-H "Content-Type: application/json" \
-H "x-goog-api-key: $COMETAPI_KEY" \
-d '{
"contents": [{"parts": [{"text": "Solve this logic puzzle step by step."}]}],
"generationConfig": {
"thinkingConfig": {"thinkingBudget": 2048}
}
}'
L’utilisation de
thinkingLevel avec les modèles Gemini 2.5 (ou de thinkingBudget avec les modèles Gemini 3) peut provoquer des erreurs. Utilisez le bon paramètre pour la version de votre modèle.Streaming
UtilisezstreamGenerateContent?alt=sse comme opérateur pour recevoir des événements Server-Sent Events pendant que le modèle génère du contenu. Chaque événement SSE contient une ligne data: avec un objet JSON GenerateContentResponse.
curl "https://api.cometapi.com/v1beta/models/gemini-2.5-flash:streamGenerateContent?alt=sse" \
-H "Content-Type: application/json" \
-H "x-goog-api-key: $COMETAPI_KEY" \
--no-buffer \
-d '{
"contents": [{"parts": [{"text": "Write a short poem about the stars"}]}]
}'
Instructions système
Guidez le comportement du modèle sur l’ensemble de la conversation avecsystemInstruction :
curl "https://api.cometapi.com/v1beta/models/gemini-2.5-flash:generateContent" \
-H "Content-Type: application/json" \
-H "x-goog-api-key: $COMETAPI_KEY" \
-d '{
"contents": [{"parts": [{"text": "What is 2+2?"}]}],
"systemInstruction": {
"parts": [{"text": "You are a math tutor. Always show your work."}]
}
}'
Mode JSON
Forcez une sortie JSON structurée avecresponseMimeType. Fournissez éventuellement un responseSchema pour une validation stricte du schéma :
curl "https://api.cometapi.com/v1beta/models/gemini-2.5-flash:generateContent" \
-H "Content-Type: application/json" \
-H "x-goog-api-key: $COMETAPI_KEY" \
-d '{
"contents": [{"parts": [{"text": "List 3 planets with their distances from the sun"}]}],
"generationConfig": {
"responseMimeType": "application/json"
}
}'
Ancrage Google Search
Activez la recherche web en temps réel en ajoutant un outilgoogleSearch :
curl "https://api.cometapi.com/v1beta/models/gemini-2.5-flash:generateContent" \
-H "Content-Type: application/json" \
-H "x-goog-api-key: $COMETAPI_KEY" \
-d '{
"contents": [{"parts": [{"text": "Who won the euro 2024?"}]}],
"tools": [{"google_search": {}}]
}'
groundingMetadata avec les URL des sources et les scores de confiance.
Exemple de réponse
Une réponse typique du endpoint Gemini de CometAPI :{
"candidates": [
{
"content": {
"role": "model",
"parts": [{"text": "Hello"}]
},
"finishReason": "STOP",
"avgLogprobs": -0.0023
}
],
"usageMetadata": {
"promptTokenCount": 5,
"candidatesTokenCount": 1,
"totalTokenCount": 30,
"trafficType": "ON_DEMAND",
"thoughtsTokenCount": 24,
"promptTokensDetails": [{"modality": "TEXT", "tokenCount": 5}],
"candidatesTokensDetails": [{"modality": "TEXT", "tokenCount": 1}]
},
"modelVersion": "gemini-2.5-flash",
"createTime": "2026-03-25T04:21:43.756483Z",
"responseId": "CeynaY3LDtvG4_UP0qaCuQY"
}
Le champ
thoughtsTokenCount dans usageMetadata indique combien de tokens le modèle a consacrés à son raisonnement interne, même lorsque la sortie de réflexion n’est pas incluse dans la réponse.Principales différences avec le endpoint compatible OpenAI
| Fonctionnalité | Gemini natif (/v1beta/models/...) | Compatible OpenAI (/v1/chat/completions) |
|---|---|---|
| Contrôle de la réflexion | thinkingConfig avec thinkingLevel / thinkingBudget | Non disponible |
| Ancrage Google Search | tools: [\{"google_search": \{\}\}] | Non disponible |
| Ancrage Google Maps | tools: [\{"googleMaps": \{\}\}] | Non disponible |
| Modalité de génération d’image | responseModalities: ["IMAGE"] | Non disponible |
| En-tête d’authentification | x-goog-api-key ou Bearer | Bearer uniquement |
| Format de réponse | Format natif Gemini (candidates, parts) | Format OpenAI (choices, message) |
Autorisations
Your CometAPI key passed via the x-goog-api-key header. Bearer token authentication (Authorization: Bearer <key>) is also supported.
Paramètres de chemin
The Gemini model ID to use. See the Models page for current Gemini model IDs.
Exemple:
"gemini-2.5-flash"
The operation to perform. Use generateContent for synchronous responses, or streamGenerateContent?alt=sse for Server-Sent Events streaming.
Options disponibles:
generateContent, streamGenerateContent?alt=sse Exemple:
"generateContent"
Corps
application/json
The conversation history and current input. For single-turn queries, provide a single item. For multi-turn conversations, include all previous turns.
Show child attributes
Show child attributes
System instructions that guide the model's behavior across the entire conversation. Text only.
Show child attributes
Show child attributes
Tools the model may use to generate responses. Supports function declarations, Google Search, Google Maps, and code execution.
Show child attributes
Show child attributes
Configuration for tool usage, such as function calling mode.
Show child attributes
Show child attributes
Safety filter settings. Override default thresholds for specific harm categories.
Show child attributes
Show child attributes
Configuration for model generation behavior including temperature, output length, and response format.
Show child attributes
Show child attributes
The name of cached content to use as context. Format: cachedContents/{id}. See the Gemini context caching documentation for details.
Réponse
200 - application/json
Successful response. For streaming requests, the response is a stream of SSE events, each containing a GenerateContentResponse JSON object prefixed with data:.
The generated response candidates.
Show child attributes
Show child attributes
Feedback on the prompt, including safety blocking information.
Show child attributes
Show child attributes
Token usage statistics for the request.
Show child attributes
Show child attributes
The model version that generated this response.
The timestamp when this response was created (ISO 8601 format).
Unique identifier for this response.
⌘I
from google import genai
client = genai.Client(
api_key="<COMETAPI_KEY>",
http_options={"api_version": "v1beta", "base_url": "https://api.cometapi.com"},
)
response = client.models.generate_content(
model="gemini-2.5-flash",
contents="Explain how AI works in a few words",
)
print(response.text){
"candidates": [
{
"content": {
"role": "<string>",
"parts": [
{
"text": "<string>",
"functionCall": {
"name": "<string>",
"args": {}
},
"inlineData": {
"mimeType": "<string>",
"data": "<string>"
},
"thought": true
}
]
},
"finishReason": "STOP",
"safetyRatings": [
{
"category": "<string>",
"probability": "<string>",
"blocked": true
}
],
"citationMetadata": {
"citationSources": [
{
"startIndex": 123,
"endIndex": 123,
"uri": "<string>",
"license": "<string>"
}
]
},
"tokenCount": 123,
"avgLogprobs": 123,
"groundingMetadata": {
"groundingChunks": [
{
"web": {
"uri": "<string>",
"title": "<string>"
}
}
],
"groundingSupports": [
{
"groundingChunkIndices": [
123
],
"confidenceScores": [
123
],
"segment": {
"startIndex": 123,
"endIndex": 123,
"text": "<string>"
}
}
],
"webSearchQueries": [
"<string>"
]
},
"index": 123
}
],
"promptFeedback": {
"blockReason": "SAFETY",
"safetyRatings": [
{
"category": "<string>",
"probability": "<string>",
"blocked": true
}
]
},
"usageMetadata": {
"promptTokenCount": 123,
"candidatesTokenCount": 123,
"totalTokenCount": 123,
"trafficType": "<string>",
"thoughtsTokenCount": 123,
"promptTokensDetails": [
{
"modality": "<string>",
"tokenCount": 123
}
],
"candidatesTokensDetails": [
{
"modality": "<string>",
"tokenCount": 123
}
]
},
"modelVersion": "<string>",
"createTime": "<string>",
"responseId": "<string>"
}