Skip to main content
Retrieve the list of models supported by the configured harness and the default model (if set).

​
Request

No parameters required.

​
Example Request

curl http://localhost:4000/models

​
Response

Returns a JSON object with available models.
​
models
array
required
Array of model identifiers in provider/model-name format. The specific models available depend on the configured harness and provider.Examples:
  • openai/gpt-4o
  • openai/gpt-4o-mini
  • anthropic/claude-3-5-sonnet-20241022
  • anthropic/claude-3-5-haiku-20241022
​
defaultModel
string
The default model used when a model is not specified in chat requests. Only included if:
  1. A default model is configured on the server
  2. The default model exists in the models array
If the configured default model is not in the supported models list, this field is omitted.

​
Example Response

{
  "models": [
    "openai/gpt-4o",
    "openai/gpt-4o-mini",
    "openai/o1",
    "openai/o1-mini",
    "anthropic/claude-3-5-sonnet-20241022",
    "anthropic/claude-3-5-haiku-20241022",
    "google/gemini-2.0-flash-exp",
    "deepseek/deepseek-chat"
  ],
  "defaultModel": "openai/gpt-4o-mini"
}

​
Example Response (No Default)

{
  "models": [
    "openai/gpt-4o",
    "anthropic/claude-3-5-sonnet-20241022"
  ]
}

​
Configuration

The default model is set via the DEFAULT_MODEL environment variable: 
DEFAULT_MODEL=openai/gpt-4o-mini bun run dev:server
Or when creating the app programmatically: 
import { createApp } from "./server/index";

const app = await createApp({
  defaultModel: "openai/gpt-4o-mini"
});

​
Notes

  • The available models depend on the configured harness and provider
  • Default harness uses the Zen provider which routes to multiple LLM providers
  • Custom harnesses can be configured to support different model sets
  • Model format follows the pattern: provider/model-name

Build docs developers (and LLMs) love

Get started for freeTalk to us