Models

Lists the models the key can call, with per-token pricing. Prices below are the list price in USD, charged to the wallet. Pass a model's id as the model field.

Media rows (image, audio, video) require a developer key. An agent key is pinned to one text model and never sees them.

The response's owned_by field is the model maker (e.g. Anthropic, OpenAI) for text, and the literal string "aihub" for every media row.

The ids below are the values passed as model. Copy them exactly, because they are API identifiers, not display names.

Posture states

Every model carries a posture, independent of whether it is currently reachable. A healthy route and a GA posture are two different facts.

Posture never contradicts live status. If a model is currently unreachable, that shows in its Status column above. The posture table only describes the plan for the model, not whether it will answer the next request.

Response shape

The table above is rendered from a live GET /models call. The raw JSON:

{
  "object": "list",
  "data": [
    {
      "id": "anthropic-claude-opus-4-8",
      "object": "model",
      "type": "text",
      "owned_by": "Anthropic",
      "pricing": { "prompt": "0.00000475", "completion": "0.00002375" }
    },
    {
      "id": "z-ai-glm-5-2",
      "object": "model",
      "type": "text",
      "owned_by": "Zhipu",
      "pricing": { "prompt": "0.0000008835", "completion": "0.00000285" }
    }
  ]
}

Pass ?type=image, ?type=audio, or ?type=video to list media models instead. Each row then also carries a parameters object describing that model's accepted fields (see Media). pricing is per-token in USD for text models; media pricing shape varies by modality.