> ## Documentation Index
> Fetch the complete documentation index at: https://docs.apiyi.com/llms.txt
> Use this file to discover all available pages before exploring further.
# Gemini 3.6 Flash Chat API 레퍼런스
> Gemini 3.6 Flash Chat Completions API 레퍼런스 및 대화형 플레이그라운드: OpenAI 호환, reasoning_effort 단계, 스트리밍, 함수 호출, 비전을 지원합니다.
오른쪽 플레이그라운드를 사용하십시오: `Bearer sk-your-api-key`를 **Authorization**에 넣으십시오. 기본 예제는 `reasoning_effort: "low"`를 사용합니다. 전송하여 응답을 확인하십시오.
**Thinking은 기본적으로 ON입니다**(출력으로 과금됩니다). 디버깅할 때는 low로 유지하십시오. 복잡한 작업에는 `max_tokens`가 4000+일 때 `high`로 전환하십시오. 검색 grounding, code execution 및 기타 네이티브 tools는 이 엔드포인트에서 사용할 수 없습니다 — 대신 [Native Playground](/ko/api-capabilities/gemini-3-6-flash/generate-content)을 사용하십시오. 개요 및 측정 데이터: [Gemini 3.6 Flash Overview](/ko/api-capabilities/gemini-3-6-flash/overview).
## 매개변수 빠른 참조
| 매개변수 | 유형 | 필수 | 비고 |
| ------------------ | ------ | -- | ----------------------------------------------------------------------- |
| `model` | string | ✓ | 고정값: `gemini-3.6-flash` |
| `messages` | array | ✓ | 표준 OpenAI 메시지; `content`는 멀티모달 배열일 수 있습니다(`image_url`는 data URL을 지원합니다) |
| `max_tokens` | int | | 출력 쿼터; thinking이 켜진 경우 2000 이상 권장(thinking은 출력으로 과금됨) |
| `reasoning_effort` | string | | `low` / `medium` / `high` thinking 단계, 작동 확인됨 |
| `stream` | bool | | SSE 스트리밍 |
| `response_format` | object | | `json_schema` 구조화 출력, 작동 확인됨 |
| `tools` | array | | Function-calling 도구 목록, 작동 확인됨 |
## 응답 참고사항
* 추론 사용량은 `usage.completion_tokens_details.reasoning_tokens`입니다(이 모델에 대해 보고됩니다)
* 이 엔드포인트에서는 Thinking text(reasoning\_content)가 다시 출력되지 않습니다 - 네이티브 엔드포인트의 `includeThoughts`를 사용하여 생각을 확인하십시오
* 모델 이름 `gemini-3.6-flash`를 다시 확인하십시오 - 오타가 있으면 조용한 대체 동작 대신 오류가 반환됩니다
## OpenAPI
````yaml api-reference/gemini-3-6-flash-chat-openapi-en.yaml POST /v1/chat/completions
openapi: 3.1.0
info:
title: Gemini 3.6 Flash Chat Completions API
description: >
Gemini 3.6 Flash (`gemini-3.6-flash`) — OpenAI-compatible Chat Completions
endpoint; any OpenAI SDK works by switching base_url.
- Input $1.50 / output $7.50 per 1M tokens (output includes thinking)
- Thinking is ON by default; tier it with `reasoning_effort`
(low/medium/high), usage reports reasoning_tokens
- Supports streaming, function calling, response_format json_schema, and
vision (data URL)
- For Search grounding / code execution and other advanced tools, use the
native Gemini endpoint
**Authentication**: header `Authorization: Bearer YOUR_API_KEY`
**Get an API Key**: create a token in the APIYI console at
`api.apiyi.com/token`
version: 1.0.0
servers:
- url: https://api.apiyi.com
description: Primary endpoint
- url: https://vip.apiyi.com
description: Backup endpoint
security:
- bearerAuth: []
paths:
/v1/chat/completions:
post:
tags:
- Text Generation
summary: 'Chat completion: Gemini 3.6 Flash (OpenAI-compatible)'
description: |
Chat completion with `gemini-3.6-flash`, fully OpenAI-compatible.
operationId: chatgemini_3_6_flash_en
requestBody:
required: true
content:
application/json:
schema:
$ref: '#/components/schemas/ChatRequest'
example:
model: gemini-3.6-flash
messages:
- role: user
content: Introduce yourself in one sentence
max_tokens: 2000
reasoning_effort: low
responses:
'200':
description: Success
content:
application/json:
schema:
$ref: '#/components/schemas/ChatResponse'
'401':
description: Invalid API Key
'429':
description: Rate limit exceeded
security:
- bearerAuth: []
components:
schemas:
ChatRequest:
type: object
required:
- model
- messages
properties:
model:
type: string
description: 'Fixed: gemini-3.6-flash'
messages:
type: array
description: Standard OpenAI messages array
items:
type: object
properties:
role:
type: string
enum:
- system
- user
- assistant
- tool
content:
description: String or multimodal array (text / image_url)
max_tokens:
type: integer
description: >-
Output quota; 2000+ recommended with thinking on (thinking bills as
output)
temperature:
type: number
stream:
type: boolean
description: SSE streaming
reasoning_effort:
type: string
enum:
- low
- medium
- high
description: Thinking tier
response_format:
type: object
description: Structured output, json_schema supported
tools:
type: array
description: Function-calling tool list
items:
type: object
ChatResponse:
type: object
properties:
id:
type: string
model:
type: string
choices:
type: array
items:
type: object
properties:
message:
type: object
properties:
role:
type: string
content:
type: string
tool_calls:
type: array
items:
type: object
finish_reason:
type: string
usage:
type: object
description: Usage stats
properties:
prompt_tokens:
type: integer
completion_tokens:
type: integer
total_tokens:
type: integer
securitySchemes:
bearerAuth:
type: http
scheme: bearer
description: 'Add Authorization: Bearer YOUR_API_KEY header'
````