Skip to main content
Retrieve all provider endpoints available for a specific model. Use this endpoint to inspect per-provider pricing, context length, quantization, and capability support.

​
Endpoint

GET /api/v1/models/{author}/{slug}/endpoints

​
Path Parameters

ParameterDescription
authorModel author (e.g. openai, anthropic, google)
slugModel slug (e.g. gpt-5.4, claude-sonnet-4.6)

​
Example Request

curl "https://api.arouter.ai/api/v1/models/openai/gpt-5.4/endpoints" \
  -H "Authorization: Bearer lr_live_xxxx"

​
Example Response

{
  "data": [
    {
      "provider_id": "OpenAI",
      "provider_name": "OpenAI",
      "model": "openai/gpt-5.4",
      "context_length": 128000,
      "max_completion_tokens": 16384,
      "pricing": {
        "prompt": 0.0000025,
        "completion": 0.0000100,
        "image": 0.0,
        "request": 0.0
      },
      "quantization": "fp16",
      "is_zdr": true,
      "supported_parameters": ["tools", "response_format", "stream", "max_tokens"]
    },
    {
      "provider_id": "Azure",
      "provider_name": "Azure OpenAI",
      "model": "openai/gpt-5.4",
      "context_length": 128000,
      "max_completion_tokens": 16384,
      "pricing": {
        "prompt": 0.0000025,
        "completion": 0.0000100
      },
      "quantization": "fp16",
      "is_zdr": true,
      "supported_parameters": ["tools", "response_format", "stream", "max_tokens"]
    }
  ]
}

​
Response Fields

FieldTypeDescription
provider_idstringProvider ID for use in routing
provider_namestringHuman-readable provider name
modelstringFull model ID
context_lengthintegerMaximum input context in tokens
max_completion_tokensintegerMaximum output tokens
pricing.promptnumberCost per prompt token in USD
pricing.completionnumberCost per completion token in USD
quantizationstringWeight quantization: fp32, fp16, bf16, fp8, int8, int4
is_zdrbooleanWhether this endpoint supports Zero Data Retention
supported_parametersstring[]Parameters this endpoint supports

​
Notes

  • Use provider IDs from this endpoint in provider.order, provider.only, and provider.ignore
  • Pricing is per token (not per 1k tokens)
  • See Provider Routing for routing by endpoint capability