Skip to main content
POST
Google
Google Gemini models cover reasoning, coding, multimodal input, and agent workflows. The playground on this page uses Gemini Native, which every Gemini model supports. Pick the model in the model path parameter.

Models

  • gemini-4-argon — Gemini 4 Argon, a frontier reasoning model with 1M context for long-horizon coding and knowledge work.
  • gemini-3.8-flash — Gemini 3.8 Flash, for software engineering, autonomous agents, and enterprise workflows.
  • gemini-3.7-flash — Gemini 3.7 Flash, for agentic coding, multimodal reasoning, and tool use.
  • gemini-3.5-flash — Gemini 3.5 Flash.
  • gemini-3.1-pro-preview — Gemini 3.1 Pro Preview.
  • gemini-3-pro-preview — Gemini 3 Pro Preview.
  • gemini-3-flash-preview — Gemini 3 Flash Preview.
Every Gemini model supports Gemini Native and Chat Completions.

Examples

  • Basic request — generateContent with contents and generationConfig.
Set method to streamGenerateContent for streaming. To call a model with the Chat Completions format, send the same model id in the model body field to Chat Completions.

Notes

  • In Gemini Native, the model id goes in the URL path, not the request body.
  • See LLM Models for every model and the formats it supports.

Authorizations

Authorization
string
header
required

All API endpoints require Bearer Token authentication.

Get your API Key:

Visit the API Key Management Page to get your API Key

Add it to the request header:

Path Parameters

model
enum<string>
default:gemini-3-flash-preview
required

Gemini model id, sent in the URL path.

Available options:
gemini-4-argon,
gemini-3.8-flash,
gemini-3.7-flash,
gemini-3.5-flash,
gemini-3.1-pro-preview,
gemini-3-pro-preview,
gemini-3-flash-preview
method
enum<string>
required

Request method. Use generateContent for complete JSON responses and streamGenerateContent for streaming generation.

Available options:
generateContent,
streamGenerateContent
Example:

"generateContent"

Body

application/json
contents
object[]
required

Conversation contents with roles and parts. Each content object may include a role (user or model) and a parts array containing text or inline data.

generationConfig
object

Generation configuration, including temperature, topP, topK, maxOutputTokens, and stopSequences.

safetySettings
object[]

Safety settings.

Response

Content generated

candidates
object[]
promptFeedback
object