Retrieve a model

Request

This endpoint takes no request body.

curl https://api.larasrouter.com/v1/models/deepseek-flash \
  -H "Authorization: Bearer $LARAS_API_KEY"

import os
from openai import OpenAI

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

print(client.models.retrieve("deepseek-flash"))

import OpenAI from "openai";

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

console.log(await client.models.retrieve("deepseek-flash"));

Response body

• idstring

  Model ID to send as model, e.g. deepseek-flash.

• objectstring

  Always model.

• createdinteger

  Unix timestamp in seconds.

• owned_bystring

  Always larasrouter.

• context_lengthinteger

  Context window in tokens, shared by the prompt and the output.

• max_output_tokensinteger

  The largest max_tokens (or max_completion_tokens) accepted. A larger value is rejected with 400 max_tokens_exceeded.

• default_output_tokensinteger

  The output cap applied when you omit max_tokens.

• capabilitiesobject

  What the gateway accepts for this model. false means a request that uses the feature is rejected.

  • streamingboolean

    stream: true (Server-Sent Events).

  • toolsboolean

    tools, tool_choice, and parallel_tool_calls.

  • json_outputboolean

    response_format: {"type": "json_object"}.

  • thinkingboolean

    Reasoning is returned in reasoning_content and billed as output.

  • visionboolean

    Image input. While false, a message with an image part is rejected with 400 unsupported_content.

• pricingobject

  The fixed price, in USD per unit_tokens tokens, as exact decimal strings (at least three decimals). The names differ from per-token fields such as OpenRouter’s prompt and completion so the units cannot be mixed up.

  • versionstring

    The price version, e.g. deepseek-flash-2026-09-v1. A price change is a new version.

  • currencystring

    Always USD.

  • unit_tokensinteger

    Tokens each price covers: 1000000.

  • input_cache_hitstring

    Price of input tokens served from the provider’s cache.

  • inputstring

    Price of input tokens that miss the cache.

  • outputstring

    Price of output tokens, reasoning included.

{
  "id": "deepseek-flash",
  "object": "model",
  "created": 1760241600,
  "owned_by": "larasrouter",
  "context_length": 1000000,
  "max_output_tokens": 384000,
  "default_output_tokens": 8192,
  "capabilities": {
    "streaming": true,
    "tools": true,
    "json_output": true,
    "thinking": true,
    "vision": false
  },
  "pricing": {
    "version": "deepseek-flash-2026-09-v1",
    "currency": "USD",
    "unit_tokens": 1000000,
    "input_cache_hit": "0.010",
    "input": "0.400",
    "output": "1.600"
  }
}

id, object, created, and owned_by are the standard OpenAI model fields, so OpenAI SDKs parse the object as usual. The other fields are Laras Router extensions: the limits the gateway enforces, the features it accepts, and the price it charges, read from the same catalogue that prices your requests.

Responses