Create chat completion

Request body

• modelstringrequired

  Model ID. The only accepted value is deepseek-flash.

• messagesarrayrequired

  The conversation so far, oldest first.

  • rolestringrequired

    system, user, assistant, or tool.

  • contentstring | array | nullrequired

    The message text, or an array of {"type": "text"} parts. Image parts are rejected with 400 unsupported_content. May be null on an assistant message that only carries tool_calls.

  • tool_callsarray

    On an assistant message: the tool calls the model made, sent back unchanged.

  • tool_call_idstring

    On a tool message: the id of the tool call this result answers.

• max_tokensinteger

  Output cap, 1 to 384,000. Defaults to 8,192 when omitted. A larger value is rejected with 400 max_tokens_exceeded, never silently reduced. Reasoning tokens count toward it.

• max_completion_tokensinteger

  Alias of max_tokens. If you send both, the smaller value applies.

• streamboolean

  Send the response as Server-Sent Events. Defaults to false.

• stream_optionsobject

  Streaming options.

  • include_usageboolean

    Send a final chunk with token usage before [DONE].

• toolsarray

  Function definitions the model may call: {"type": "function", "function": {name, description, parameters}}.

• tool_choicestring | object

  auto, none, required, or a specific function.

• response_formatobject

  {"type": "json_object"} makes the model return a valid JSON object.

• temperature, top_p, stop, presence_penalty, frequency_penalty—

  Sampling parameters, forwarded to the model unchanged.

Example request

curl https://api.larasrouter.com/v1/chat/completions \
  -H "Authorization: Bearer $LARAS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-flash",
    "messages": [
      {"role": "system", "content": "You are a concise assistant."},
      {"role": "user", "content": "Name three uses of a hash map."}
    ],
    "max_tokens": 1024
  }'

import os
from openai import OpenAI

client = OpenAI(
    base_url="https://api.larasrouter.com/v1",
    api_key=os.environ["LARAS_API_KEY"],
)

response = client.chat.completions.create(
    model="deepseek-flash",
    messages=[
        {"role": "system", "content": "You are a concise assistant."},
        {"role": "user", "content": "Name three uses of a hash map."},
    ],
    max_tokens=1024,
)
print(response.choices[0].message.content)

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.larasrouter.com/v1",
  apiKey: process.env.LARAS_API_KEY,
});

const response = await client.chat.completions.create({
  model: "deepseek-flash",
  messages: [
    { role: "system", content: "You are a concise assistant." },
    { role: "user", content: "Name three uses of a hash map." },
  ],
  max_tokens: 1024,
});
console.log(response.choices[0].message.content);

Response body

• idstring

  Completion ID.

• objectstring

  chat.completion, or chat.completion.chunk for stream chunks.

• createdinteger

  Unix timestamp in seconds.

• modelstring

  The model that answered.

• choicesarray

  One choice.

  • message.contentstring | null

    The answer text.

  • message.reasoning_contentstring | null

    The model’s reasoning, when it produced any. Billed as output.

  • message.tool_callsarray

    Tool calls the model wants you to run.

  • finish_reasonstring

    stop, length (output cap reached), or tool_calls.

• usageobject

  Token counts the charge is computed from.

  • prompt_tokensinteger

    All input tokens.

  • prompt_cache_hit_tokensinteger

    Input tokens served from cache, billed at $0.010 / 1M.

  • prompt_cache_miss_tokensinteger

    Input tokens not in cache, billed at $0.400 / 1M.

  • completion_tokensinteger

    Output tokens, reasoning included, billed at $1.600 / 1M.

  • total_tokensinteger

    Input plus output.

{
  "id": "chatcmpl-…",
  "object": "chat.completion",
  "created": 1760241600,
  "model": "deepseek-flash",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "1. Caching results by key…",
        "reasoning_content": "The user wants three uses…"
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 24,
    "completion_tokens": 180,
    "total_tokens": 204,
    "prompt_cache_hit_tokens": 0,
    "prompt_cache_miss_tokens": 24
  }
}

Streaming

With "stream": true the response is a stream of chat.completion.chunk events, ending with data: [DONE]. Set stream_options.include_usage to receive token usage in a final chunk. See Streaming.

data: {"id":"chatcmpl-…","object":"chat.completion.chunk","choices":[{"index":0,"delta":{"role":"assistant","content":""}}]}

data: {"id":"chatcmpl-…","object":"chat.completion.chunk","choices":[{"index":0,"delta":{"content":"1. Caching"}}]}

data: {"id":"chatcmpl-…","object":"chat.completion.chunk","choices":[{"index":0,"delta":{},"finish_reason":"stop"}]}

data: {"id":"chatcmpl-…","object":"chat.completion.chunk","choices":[],"usage":{"prompt_tokens":24,"completion_tokens":180,"total_tokens":204,"prompt_cache_hit_tokens":0,"prompt_cache_miss_tokens":24}}

data: [DONE]

Responses