---
title: "Retrieve Model"
method: GET
path: "/api/v1/models/{model_id}"
tags: ["models"]
---

# Retrieve Model

`GET /api/v1/models/{model_id}`

OpenAI-style single model retrieval.

Supports both base models and LoRAs:
- Base model: /models/meta-llama/Llama-3.1-8B
- LoRA: /models/meta-llama/Llama-3.1-8B:my-adapter

Note: Some base models contain colons (e.g., OpenRouter :free variants).
We first try exact match as base model, then fall back to LoRA parsing.

Returns 404 for unauthorized access to avoid leaking model existence.

## Path parameters

- `model_id` string, required — Model id (base_model or base_model:lora_id); may contain '/'

## Headers

- `authorization` string, nullable
- `x-api-key` string, nullable
- `X-Prime-Team-Id` string, nullable

## Response `200`

Successful Response

- OpenAIModelObject — OpenAI-compatible model object schema.
  - `id` string, required
  - `object` string
  - `owned_by` string, nullable
  - `created` integer, nullable
  - `pricing` PricingInfo
    - `input_usd_per_mtok` number, nullable
    - `output_usd_per_mtok` number, nullable
    - `effective_at` string, nullable

## Other responses

- `422` — Validation Error

---

[API](https://skmtc.dev/pinference/apis/pi-inference.md) · [All operations](https://skmtc.dev/pinference/apis/pi-inference/llms.txt) · [OpenAPI document](https://skmtc-service-production.skmtc.workers.dev/v1/apis/pinference/pi-inference/revisions/707211c46ce9/schema)