List models

GET /v1/models — the catalogue, scoped to what your key can actually call.

GEThttps://iranrouter.com/v1/models

The catalogue, scoped to what this key can call. Authenticated, because the list depends on the key.

Shell
curl https://iranrouter.com/v1/models -H "Authorization: Bearer ir-..."
JSON
{
  "object": "list",
  "data": [
    {
      "id": "openai/gpt-4o-mini",
      "object": "model",
      "display_name": "GPT-4o mini",
      "author": { "slug": "openai", "name": "OpenAI" },
      "architecture": { "input_modalities": ["text", "image"], "output_modalities": ["text"] },
      "endpoint": "/v1/chat/completions",
      "context_window": 128000,
      "max_output_tokens": 16384,
      "capabilities": { "tools": true, "vision": true, "json_schema": true },
      "supported_parameters": ["tools", "vision", "json_schema"],
      "pricing": { "input_rial_per_mtok": 1500000, "output_rial_per_mtok": 6000000, "currency": "IRR" },
      "billed_from": "plan"
    }
  ]
}
FieldMeaning
idExactly the string you put in the model field.
architectureWhat the model takes and makes: text, image, audio, video, file, embeddings, transcription, speech, rerank.
endpointThe path this model answers on. Do not call an embedding model on chat/completions. null means a family not served yet.
context_windowThe input + output token ceiling.
capabilitiesDeclared capabilities; supported_parameters is the list of the true keys.
pricingInput and output rates in Rial per million tokens. The other rates (cache, reasoning, image, audio, conditional prices) are on the model's page.
billed_from'plan' draws the allowance, 'credit' bills the wallet.
deprecated / sunset_at / successorOnly on a deprecated model: it works until sunset_at, then resolves to successor.
GEThttps://iranrouter.com/v1/models/{id}

One model, in the same shape as a list row. The id contains a slash; do not encode the path: GET https://iranrouter.com/v1/models/openai/gpt-4o-mini. A model outside this key's list is a 404.

The public catalogue with Rial pricing and benchmarks is at the models page.