Models

The catalog is public. Agents can check models and rates before they spend anything.

List models

The catalog is public, so an agent can discover models and rates before it spends anything.

curl https://inferlane.xyz/v1/models

{
  "object": "list",
  "data": [
    {
      "id": "meta-llama/llama-3.3-70b-instruct",
      "object": "model",
      "owned_by": "meta-llama",
      "type": "chat",
      "context_length": 131072,
      "input_usd_per_mtok": 0.10,
      "output_usd_per_mtok": 0.32,
      "min_output_tokens": 1,
      "released_at": "2024-12-06"
    }
  ]
}

The shape follows the OpenAI list format, so model pickers in existing SDKs and tools work unchanged. The same catalog is browsable on the models page.

Fields

• type: chat or embeddings, which decides the endpoint to call.
• input_usd_per_mtok / output_usd_per_mtok: the posted rate per 1M tokens. This is exactly what you are charged, settled on Robinhood Chain.
• context_length: the window. It also caps the pay-per-call ceiling.
• min_output_tokens: reasoning models think before answering. A max_tokens below this is refused up front rather than billed for an empty reply.

Model ids

Ids are family/model, for example meta-llama/llama-3.3-70b-instruct. Ids are stable once published; a retired model answers 404 model_not_found instead of silently switching to something else.