gemini-flash-lite Chat API

API health

24h health: Normal
24h average time: —

Healthy / normal
Available
Limited
Mostly down

Normal

Last 24 hours

Last 30 days

gemini-flash-lite is the lighter model of the EveryInfra AI API. Set "model": "gemini-flash-lite" to use it; requests without a model field use the default model gemini-3.6-flash. The API is compatible with OpenAI Chat Completions, so an existing OpenAI SDK only needs a new base_url and api_key.

Free for accounts that have topped up any amount; calls are not billed. Accounts that have never topped up cannot call it.

Request parameters

Name	In	Type	Required	Default	Description
`Authorization`	header	`string`	Yes	-	Bearer followed by your API key. Create it in the console; never put it in a URL or a code repository.
`model`	body	`string`	No	`gemini-3.6-flash`	Model name. Without it the default model gemini-3.6-flash is used; GET /api/v1/models lists the models and needs no key.
`messages`	body	`array`	Yes	-	Conversation messages in OpenAI format (role + content). All messages together may hold up to 45,000 characters; longer input returns 422, so split it into several requests.
`stream`	body	`boolean`	No	`false`	When true, the reply arrives as SSE and ends with data: [DONE]. Default false.
`stream_options`	body	`object`	No	-	Streaming options in OpenAI format; with include_usage set to true the last chunk carries token usage.

Other OpenAI Chat Completions parameters (temperature, max_tokens, response_format and so on) are passed to the model unchanged.

Code samples

bash

curl https://api.everyinfra.com/api/v1/chat/completions \
  -H "Authorization: Bearer $EVERYINFRA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "gemini-flash-lite",
  "messages": [
    {
      "role": "user",
      "content": "Summarize this review in one sentence: ..."
    }
  ]
}'

Samples read the key from the EVERYINFRA_API_KEY environment variable; copied code never contains your key.

Response example

Field preview

Example structure, not a live call: choices[0].message.content is the model's reply and usage is the token count. The format is the same as OpenAI's.

json

{
  "id": "chatcmpl-…",
  "object": "chat.completion",
  "created": …,
  "model": "gemini-flash-lite",
  "choices": [
    {
      "index": …,
      "message": {
        "role": "assistant",
        "content": "…"
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": …,
    "completion_tokens": …,
    "total_tokens": …
  }
}

Rate limits

Tiers follow the account's total top-ups (bonus credit and redeem codes do not count), across all keys of the account. These are limits, not promised throughput.

Topped up any amount: 50 requests per minute, up to 5 in flight
Topped up $100+: 500 requests per minute, up to 50 in flight
Topped up $500+: 5,000 requests per minute, up to 500 in flight

Errors and billing

401 unauthorized: The key is missing or invalid.
403 ai_topup_required: The account has never topped up. Top up any amount to start.
422 unknown_model / missing_param / input_too_large: Wrong model name (the error suggests the closest one), messages missing, or input too long. Fix the request and send it again.
429 rate_limited: Over this tier's requests per minute. Wait for Retry-After and try again.
503: The model is temporarily unavailable. Try again later.

Errors