> ## 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 3.5 Flash-Lite Native API Reference
> Gemini 3.5 Flash-Lite native generateContent API reference and interactive playground: official request format, zero thinking by default, Search grounding and other tools supported.
Use the playground on the right: put `sk-your-api-key` (your APIYI token — no Google Key needed) in **x-goog-api-key** and hit send — no thinking by default, so responses are fast.
**No thinking output by default.** For deep reasoning pass `thinkingConfig: {"thinkingLevel": "high"}` (only the high tier reliably triggers thinking in our tests). Search grounding, code execution, URL context, and Maps tools work only on this native endpoint. Capabilities and pricing: [Gemini 3.5 Flash-Lite Overview](/en/api-capabilities/gemini-3-5-flash-lite/overview).
* For streaming, change the URL to `:streamGenerateContent?alt=sse` (the playground demonstrates non-streaming)
* Code execution works, but `executableCode` / `codeExecutionResult` fields are currently not echoed — results appear in the text body
* `:countTokens` and the explicit cache API are not yet enabled; Computer Use is not supported by this model officially
## Parameter quick reference
| Parameter | Type | Required | Notes |
| ------------------------------------------------------ | ------ | -------- | --------------------------------------------------------------------------------------------- |
| `contents` | array | ✓ | Conversation contents; `parts` can mix `text` and `inlineData` (base64 image/PDF/audio/video) |
| `systemInstruction` | object | | System instruction |
| `generationConfig.thinkingConfig.thinkingLevel` | string | | No thinking by default; `high` measured at \~1000 thinking tokens |
| `generationConfig.thinkingConfig.includeThoughts` | bool | | Echo thought parts (pair with high tier) |
| `generationConfig.responseMimeType` + `responseSchema` | | | Structured output (JSON Schema) |
| `tools` | array | | `google_search` / `url_context` / `codeExecution` / `google_maps` / `functionDeclarations` |
## Response notes
* Thinking spend is in `usageMetadata.thoughtsTokenCount` (0 by default)
* Grounding sources appear in `candidates[0].groundingMetadata`; URL context in `urlContextMetadata`
* No implicit-cache hits observed for this model in our tests — don't build cost models on them
## OpenAPI
````yaml api-reference/gemini-3-5-flash-lite-native-openapi-en.yaml POST /v1beta/models/gemini-3.5-flash-lite:generateContent
openapi: 3.1.0
info:
title: Gemini 3.5 Flash-Lite Native generateContent API
description: >
Gemini 3.5 Flash-Lite (`gemini-3.5-flash-lite`) — native Gemini
generateContent endpoint. Call it directly with your APIYI token; no Google
API Key needed.
- Input $0.30 / output $2.50 per 1M tokens (output includes thinking), 1M
context / 64K output
- Thinking is OFF by default for minimal latency; pass
`thinkingConfig.thinkingLevel: "high"` when you need deep reasoning (in our
tests only the high tier reliably triggers thinking)
- Advanced tools (Search/Maps grounding, URL context, code execution) are
exclusive to this native format
**Authentication**: header `x-goog-api-key: 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:
- geminiKey: []
paths:
/v1beta/models/gemini-3.5-flash-lite:generateContent:
post:
tags:
- Text Generation
summary: 'Text generation: Gemini 3.5 Flash-Lite (native Gemini format)'
description: >
Generate content with `gemini-3.5-flash-lite`, fully compatible with
Google's official request format. For streaming use
`:streamGenerateContent?alt=sse`.
operationId: generategemini_3_5_flash_lite_en
requestBody:
required: true
content:
application/json:
schema:
$ref: '#/components/schemas/GenerateContentRequest'
example:
contents:
- role: user
parts:
- text: Introduce yourself in one sentence
responses:
'200':
description: Success
content:
application/json:
schema:
$ref: '#/components/schemas/GenerateContentResponse'
'401':
description: Invalid API Key
'429':
description: Rate limit exceeded
security:
- geminiKey: []
components:
schemas:
GenerateContentRequest:
type: object
required:
- contents
properties:
contents:
type: array
description: Conversation contents; parts can mix modalities
items:
type: object
properties:
role:
type: string
enum:
- user
- model
parts:
type: array
items:
type: object
properties:
text:
type: string
inlineData:
type: object
description: Inline base64 media (image/PDF/audio/video)
properties:
mimeType:
type: string
data:
type: string
systemInstruction:
type: object
description: System instruction
properties:
parts:
type: array
items:
type: object
properties:
text:
type: string
generationConfig:
type: object
description: Generation config
properties:
temperature:
type: number
maxOutputTokens:
type: integer
responseMimeType:
type: string
description: Set to application/json for structured output
responseSchema:
type: object
description: JSON Schema for structured output
thinkingConfig:
type: object
description: Thinking config (see overview for measured tiers)
properties:
thinkingLevel:
type: string
enum:
- minimal
- low
- medium
- high
thinkingBudget:
type: integer
includeThoughts:
type: boolean
tools:
type: array
description: >-
Tools: google_search / url_context / codeExecution /
functionDeclarations, etc.
items:
type: object
GenerateContentResponse:
type: object
properties:
candidates:
type: array
items:
type: object
properties:
content:
type: object
properties:
role:
type: string
parts:
type: array
items:
type: object
finishReason:
type: string
groundingMetadata:
type: object
description: Returned when grounding tools are enabled
usageMetadata:
type: object
description: Usage; thoughtsTokenCount is the thinking spend
properties:
promptTokenCount:
type: integer
candidatesTokenCount:
type: integer
thoughtsTokenCount:
type: integer
totalTokenCount:
type: integer
modelVersion:
type: string
securitySchemes:
geminiKey:
type: apiKey
in: header
name: x-goog-api-key
description: APIYI token, the sk- prefixed key
````