Skip to main content
GET
/
models
List all gateway models
curl --request GET \
  --url https://api.ai-stats.phaseo.app/v1/models \
  --header 'Authorization: Bearer <token>'
{
  "ok": true,
  "limit": 123,
  "offset": 123,
  "models": [
    {
      "model_id": "<string>",
      "name": "<string>",
      "release_date": "2023-12-25",
      "announcement_date": "2023-12-25",
      "status": "<string>",
      "organisation": "ai21",
      "aliases": [
        "<string>"
      ],
      "endpoints": [
        "chat.completions"
      ],
      "providers": [
        {
          "api_provider_id": "<string>",
          "api_provider_name": "<string>",
          "link": "<string>",
          "country_code": "<string>",
          "endpoint": "chat.completions",
          "provider_model_slug": "<string>",
          "is_active_gateway": true,
          "input_modalities": [
            "<string>"
          ],
          "output_modalities": [
            "<string>"
          ],
          "effective_from": "2023-11-07T05:31:56Z",
          "effective_to": "2023-11-07T05:31:56Z"
        }
      ],
      "input_types": [
        "<string>"
      ],
      "output_types": [
        "<string>"
      ]
    }
  ],
  "total": 123
}
Each catalogue entry now exposes two optional sections so you can tailor the payload to what your client needs:
  • data – details sourced from the AI Stats dataset such as supported input/output modalities.
  • gateway – live routing metadata showing the endpoints and providers that are currently active through the gateway.
Pass include_data=false and/or include_gateway=false to omit either section (both default to true). All of the existing filters—provider, organisation, include_endpoints, exclude_endpoints, input_types, and output_types—continue to work exactly the same, and the top-level model metadata (name, identifiers, statuses, aliases, organisation) stays unchanged for backwards compatibility.

Authorizations

​
Authorization
string
header
required

Bearer authentication header of the form Bearer <token>, where <token> is your auth token.

Query Parameters

​
provider

Filter results to models served by one or more provider identifiers.

​
limit
integer

Maximum number of models to return (default 50).

Required range: 1 <= x <= 250
​
offset
integer

Number of models to skip before starting the page.

Required range: x >= 0
​
organisation

Restrict results to models associated with one or more organisation identifiers. Organisation identifier.

Available options:
ai21,
amazon,
anthropic,
baidu,
black-forest-labs,
bytedance,
cohere,
deepseek,
eleven-labs,
google,
ibm,
inclusionai,
lg,
meta,
microsoft,
minimax,
mistral,
moonshotai,
nous,
nvidia,
openai,
perplexity,
qwen,
suno,
x-ai,
z-ai
​
include_endpoints

Only return models that support at least one of the specified gateway endpoints.

Available options:
chat.completions,
images.generations,
audio.speech,
audio.transcription,
moderations,
video.generation,
embeddings,
batch
​
exclude_endpoints

Exclude models that support any of the specified gateway endpoints.

Available options:
chat.completions,
images.generations,
audio.speech,
audio.transcription,
moderations,
video.generation,
embeddings,
batch
​
input_types

Only return models that advertise support for at least one of the requested input content types.

​
output_types

Only return models that advertise support for at least one of the requested output content types.

​
include_rumoured
boolean
default:true

Whether to include models marked as rumoured in the response (default true).

​
include_deprecated
boolean
default:true

Whether to include models marked as deprecated in the response (default true).

​
include_retired
boolean
default:true

Whether to include models marked as retired in the response (default true).

Response

Paginated model catalogue.

​
ok
boolean
required
​
limit
integer
required
​
offset
integer
required
​
models
object[]
required
​
total
integer
required

Total number of models matching the filter.