API ReferenceModels

OpenRouter Model Listing

Return model metadata in a denser per-token listing format

Use GET /v1/models?format=openrouter when you need a denser per-model listing with exact per-token prices, exact context length in tokens, quantization, Hugging Face checkpoint ID, datacenter country codes, and a supported-features enum. The plain GET /v1/models endpoint continues to return the standard EmpirioLabs catalog shape.

The OpenRouter view returns only models flagged for this listing and includes the following fields per model: id, hugging_face_id, name, created, input_modalities, output_modalities, quantization, context_length, max_output_length, pricing, supported_sampling_parameters, supported_features, is_ready, openrouter, and datacenters.

$curl "https://api.empiriolabs.ai/v1/models?format=openrouter"

Response shape

1{
2  "data": [
3    {
4      "id": "qwen3-5-9b",
5      "hugging_face_id": "Qwen/Qwen3.5-9B",
6      "name": "Qwen3.5 9B",
7      "created": 1778976000,
8      "input_modalities": ["text", "image", "video"],
9      "output_modalities": ["text"],
10      "quantization": "fp8",
11      "context_length": 262144,
12      "max_output_length": 32768,
13      "pricing": {
14        "prompt": "0.00000009",
15        "completion": "0.00000013",
16        "image": "0",
17        "request": "0",
18        "input_cache_read": "0.000000045"
19      },
20      "supported_sampling_parameters": [
21        "temperature",
22        "top_p",
23        "top_k",
24        "min_p",
25        "frequency_penalty",
26        "presence_penalty",
27        "repetition_penalty",
28        "stop",
29        "seed",
30        "max_tokens",
31        "logit_bias"
32      ],
33      "supported_features": [
34        "tools",
35        "json_mode",
36        "structured_outputs",
37        "reasoning"
38      ],
39      "is_ready": true,
40      "openrouter": {
41        "slug": "qwen/qwen3.5-9b"
42      },
43      "datacenters": [
44        {
45          "country_code": "JP"
46        }
47      ]
48    }
49  ]
50}

Field notes

FieldMeaning
idModel identifier used when calling the API.
hugging_face_idHugging Face checkpoint identifier when the model is served from Hugging Face weights.
context_lengthMaximum total context length accepted by the model, in tokens.
max_output_lengthMaximum generated-token budget allowed for one request.
pricingUSD prices as strings, expressed per token unless the field name says otherwise.
input_cache_readCached input-token price when cached-token accounting is available.
supported_sampling_parametersSampling controls accepted by the model.
supported_featuresHigher-level capabilities: tools, json_mode, structured_outputs, reasoning, logprobs, web_search.
is_readyWhether the model should be treated as ready for listing.
openrouter.slugSuggested public slug for this model.
datacenters[].country_codeISO 3166-1 alpha-2 code for the current serving region.