GET /models

The model catalog with regions, prices in PnP, context and capabilities. No key needed.

GET/api/v1/models

Lists every model you can call now. No API key is needed; the list is cached for 60 seconds and limited to 120 requests per minute per IP address.

Response

{
  "data": [
    {
      "id": "google/gemma-3-4b-it",
      "name": "Google: Gemma 3 4B",
      "description": "…",
      "context_length": 131072,
      "architecture": { "input_modalities": ["text", "image"], "output_modalities": ["text"], "modality": "text+image->text" },
      "pricing": { "prompt": "0.0000000176", "completion": "0.0000000704", "request": "0", "input_cache_read": "0" },
      "pricing_currency": "PnP",
      "top_provider": { "context_length": 131072, "max_completion_tokens": 8192 },
      "supported_parameters": ["max_tokens", "temperature", "top_p", "stop", "response_format"],
      "routes": [
        {
          "provider": "global",
          "pricing": { "prompt": "0.0000000176", "completion": "0.0000000704", "request": "0", "input_cache_read": "0" },
          "context_length": 131072,
          "max_completion_tokens": 8192,
          "zdr": true
        }
      ]
    }
  ]
}

Field
`id`	The id to send in `model`.
`context_length`	Input and output tokens the model can take together.
`architecture.input_modalities`	What the model accepts: `text`, `image`, `file`, `audio`, `video`.
`pricing`	PnP per token (`prompt`, `completion`, `input_cache_read`) and per request (`request`), with the markup included: the cheapest route.
`supported_parameters`	Request fields the model supports.
`routes[].provider`	The region: `global` or `ru`.
`routes[].pricing`	The route's prices.
`routes[].context_length`, `routes[].max_completion_tokens`	The route's context and output limit.
`routes[].zdr`	`true` when the route has a zero data retention endpoint, see Zero data retention.

Example

curl -s https://router.nodul.ru/api/v1/models | jq '.data[] | {id, prompt: .pricing.prompt}'