Models

GET /v1/models lists every model your key can call, in OpenAI's list format with a few extra fields per model:

curl https://inference.hindsight.vectorize.io/v1/models \
  -H "Authorization: Bearer hs_your_api_key"

{
  "object": "list",
  "data": [
    {
      "id": "hs-embed-1",
      "object": "model",
      "owned_by": "hindsight",
      "model_class": "embedding",
      "group": "embedding",
      "endpoint": "/v1/embeddings",
      "default": true,
      "pricing": { "input_per_million": "0.005" },
      "limits": { "max_inputs": 128 }
    },
    {
      "id": "hs-decide-1-flash",
      "object": "model",
      "owned_by": "hindsight",
      "model_class": "rerank",
      "group": "rerank",
      "endpoint": "/v1/rerank",
      "default": true,
      "pricing": { "input_per_million": "0.02" },
      "limits": { "max_documents": 1000, "max_tokens_per_doc": 4096 }
    }
  ]
}

Field	Meaning
id	What to send as model
group	The API that serves it: embedding, rerank or decision
model_class	What it is billed as. Decision models are billed as reranking
endpoint	Where to call it
default	true for the model a request to that endpoint gets when it names no model
pricing	USD per million tokens, the rates you are billed at. output_per_million appears only for models that produce output. null if prices can't be read at the moment
limits	What a call is held to: for embeddings max_inputs; for reranking max_documents and max_tokens_per_doc
display_name, description	Present when the model has them

The list is the same for every key and is the source of truth for what the API serves — the console's Models page is built from it. See Pricing & billing for the rates.