Models

Returns every active model, of every type. Read architecture.output_modalities to tell text models from image and embeddings models. No key is needed. Claude Code reads this endpoint for model discovery, with ?limit=1000.

Returns one model. Ids contain a slash, so write it as part of the path: /v1/models/anthropic/claude-haiku-4.5. An unknown id returns 404 model_not_found.

Example

curl https://deference.si/v1/models

import requests

models = requests.get("https://deference.si/v1/models").json()["data"]
print([model["id"] for model in models][:5])

const response = await fetch("https://deference.si/v1/models");
const { data } = await response.json();
console.log(data.slice(0, 5).map((model: { id: string }) => model.id));

Response

{
  "object": "list",
  "data": [
    {
      "id": "anthropic/claude-sonnet-5.5",
      "object": "model",
      "created": 1760000000,
      "owned_by": "anthropic",
      "display_name": "Claude Sonnet 5.5",
      "description": "A fast model for coding and agents.",
      "context_length": 200000,
      "max_output_tokens": 64000,
      "architecture": { "input_modalities": ["text", "image"], "output_modalities": ["text"] },
      "pricing": {
        "prompt": "0.000003",
        "completion": "0.000015",
        "input_cache_read": "0.0000003",
        "input_cache_write": "0.00000375",
        "request": "0",
        "image": "0"
      },
      "supported_parameters": ["tools", "tool_choice", "max_tokens", "temperature"]
    }
  ],
  "has_more": false,
  "first_id": "anthropic/claude-sonnet-5.5",
  "last_id": "anthropic/claude-sonnet-5.5"
}

Fields

idstring

The id to send as model.

display_namestring

A readable name.

descriptionstring

What the model is for.

context_lengthinteger

Context window in tokens.

max_output_tokensinteger

The largest output the model allows.

architectureobject

input_modalities and output_modalities.

pricingobject

USD per token as strings: prompt, completion, input_cache_read, input_cache_write. request and image are USD per request and per image.

supported_parametersarray

Request parameters the model accepts.