Gemini Native Format

Use the Gemini native API format for text generation, streaming output, and thinking mode

Gemini Native Format

Crazyrouter supports Google Gemini’s native API format, including the generateContent and streamGenerateContent endpoints.

Endpoints

POST /v1beta/models/{model}:generateContent
POST /v1beta/models/{model}:streamGenerateContent

Authentication

Pass the API Key via URL parameter:

?key=YOUR_API_KEY

---

Text Generation

cURL

curl "https://crazyrouter.com/v1beta/models/gemini-2.5-flash:generateContent?key=YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "contents": [
      {
        "role": "user",
        "parts": [
          {"text": "Explain what machine learning is in simple terms"}
        ]
      }
    ],
    "generationConfig": {
      "temperature": 0.7,
      "maxOutputTokens": 1024
    }
  }'

Response Format

{
  "candidates": [
    {
      "content": {
        "parts": [
          {
            "text": "Machine learning is a branch of artificial intelligence..."
          }
        ],
        "role": "model"
      },
      "finishReason": "STOP",
      "index": 0
    }
  ],
  "usageMetadata": {
    "promptTokenCount": 12,
    "candidatesTokenCount": 150,
    "totalTokenCount": 162
  }
}

---

Streaming Generation

cURL

curl "https://crazyrouter.com/v1beta/models/gemini-2.5-flash:streamGenerateContent?key=YOUR_API_KEY&alt=sse" \
  -H "Content-Type: application/json" \
  -d '{
    "contents": [
      {
        "role": "user",
        "parts": [{"text": "Write a short essay about the future of artificial intelligence"}]
      }
    ]
  }'

---

Multi-Turn Conversation

cURL

curl "https://crazyrouter.com/v1beta/models/gemini-2.5-flash:generateContent?key=YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "contents": [
      {
        "role": "user",
        "parts": [{"text": "Hi, I want to learn programming"}]
      },
      {
        "role": "model",
        "parts": [{"text": "Hello! Learning to program is a great choice. Do you have a specific direction in mind?"}]
      },
      {
        "role": "user",
        "parts": [{"text": "I want to learn Python, where should I start?"}]
      }
    ]
  }'

---

Thinking Mode

Gemini thinking models support reasoning before answering:

cURL

curl "https://crazyrouter.com/v1beta/models/gemini-2.5-flash:generateContent?key=YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "contents": [
      {
        "role": "user",
        "parts": [{"text": "Solve this equation: x^3 - 6x^2 + 11x - 6 = 0"}]
      }
    ],
    "generationConfig": {
      "thinkingConfig": {
        "thinkingBudget": 8000
      },
      "maxOutputTokens": 16000
    }
  }'

generationConfig Parameters

Parameter	Type	Description
temperature	number	Sampling temperature, 0-2
maxOutputTokens	integer	Maximum output tokens
topP	number	Nucleus sampling parameter
topK	integer	Top-K sampling
stopSequences	array	Stop sequences
responseMimeType	string	Response MIME type, e.g. application/json
responseSchema	object	JSON Schema to constrain output format
thinkingConfig	object	Thinking mode configuration

Gemini native format uses contents arrays and parts structures, which differ from OpenAI’s messages format. If you prefer the OpenAI format, you can use the Gemini OpenAI-compatible format.