Gemini Generating Content

Use the Gemini native API format through CometAPI for text generation, multimodal input, thinking/reasoning, function calling, Google Search grounding, JSON mode, and streaming.

POST /v1beta/models/{model}:{operator}

Python

Basic

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)

200

{
  "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>"
}

Overview

CometAPI supports the Gemini native API format, giving you full access to Gemini-specific features like thinking control, Google Search grounding, native image generation modalities, and more. Use this endpoint when you need capabilities not available through the OpenAI-compatible chat endpoint.

Quick Start

Replace the base URL and API key in any Gemini SDK or HTTP client:

Setting	Google Default	CometAPI
Base URL	generativelanguage.googleapis.com	api.cometapi.com
API Key	$GEMINI_API_KEY	$COMETAPI_KEY

Both x-goog-api-key and Authorization: Bearer headers are supported for authentication.

Thinking (Reasoning)

Gemini models can perform internal reasoning before generating a response. The control method depends on the model generation.

Gemini 3 (thinkingLevel)

Gemini 3 models use thinkingLevel to control reasoning depth. Available levels: 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"}
    }
  }'

Gemini 2.5 (thinkingBudget)

Gemini 2.5 models use thinkingBudget for fine-grained token-level control:

• 0 — disable thinking
• -1 — dynamic (model decides, default)
• > 0 — specific token budget (e.g., 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}
    }
  }'

Using thinkingLevel with Gemini 2.5 models (or thinkingBudget with Gemini 3 models) may cause errors. Use the correct parameter for your model version.

---

Streaming

Use streamGenerateContent?alt=sse as the operator to receive Server-Sent Events as the model generates content. Each SSE event contains a data: line with a JSON GenerateContentResponse object.

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"}]}]
  }'

---

System Instructions

Guide the model’s behavior across the entire conversation with systemInstruction:

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."}]
    }
  }'

---

JSON Mode

Force structured JSON output with responseMimeType. Optionally provide a responseSchema for strict schema validation:

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"
    }
  }'

---

Google Search Grounding

Enable real-time web search by adding a googleSearch tool:

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": {}}]
  }'

The response includes groundingMetadata with source URLs and confidence scores.

---

Response Example

A typical response from CometAPI’s Gemini endpoint:

{
  "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"
}

The thoughtsTokenCount field in usageMetadata shows how many tokens the model spent on internal reasoning, even when thinking output is not included in the response.

---

Key Differences from OpenAI-Compatible Endpoint

Feature	Gemini Native (/v1beta/models/...)	OpenAI-Compatible (/v1/chat/completions)
Thinking control	thinkingConfig with thinkingLevel / thinkingBudget	Not available
Google Search grounding	tools: [\{"google_search": \{\}\}]	Not available
Google Maps grounding	tools: [\{"googleMaps": \{\}\}]	Not available
Image generation modality	responseModalities: ["IMAGE"]	Not available
Auth header	x-goog-api-key or Bearer	Bearer only
Response format	Gemini native (candidates, parts)	OpenAI format (choices, message)

Authorizations

x-goog-api-key · string · header · required

Your CometAPI key passed via the x-goog-api-key header. Bearer token authentication (Authorization: Bearer <key>) is also supported.

Path Parameters

model · string · required

The Gemini model ID to use. See the Models page for current Gemini model IDs.

Example:

"gemini-2.5-flash"

operator · enum<string> · required

The operation to perform. Use generateContent for synchronous responses, or streamGenerateContent?alt=sse for Server-Sent Events streaming.

Available options: generateContent, streamGenerateContent?alt=sse

Example:

"generateContent"

Body

application/json

contents · object[] · required

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

systemInstruction · object

System instructions that guide the model's behavior across the entire conversation. Text only.

Show child attributes

tools · object[]

Tools the model may use to generate responses. Supports function declarations, Google Search, Google Maps, and code execution.

Show child attributes

toolConfig · object

Configuration for tool usage, such as function calling mode.

Show child attributes

safetySettings · object[]

Safety filter settings. Override default thresholds for specific harm categories.

Show child attributes

generationConfig · object

Configuration for model generation behavior including temperature, output length, and response format.

Show child attributes

cachedContent · string

The name of cached content to use as context. Format: cachedContents/{id}. See the Gemini context caching documentation for details.

Response

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:.

candidates · object[]

The generated response candidates.

Show child attributes

promptFeedback · object

Feedback on the prompt, including safety blocking information.

Show child attributes

usageMetadata · object

Token usage statistics for the request.

Show child attributes

modelVersion · string

The model version that generated this response.

createTime · string

The timestamp when this response was created (ISO 8601 format).

responseId · string

Unique identifier for this response.