---
title: "List inference instance types"
method: GET
path: "/v2/public/inference-instance-types"
tags: ["InstanceTypeService"]
---

# List inference instance types

`GET /v2/public/inference-instance-types`

## Request

- Base URL: `https://api.together.ai`
- URL: `https://api.together.ai/v2/public/inference-instance-types`
- Auth: HTTP bearer

Lists hardware instance types currently available to inference deployments, including GPU resources, pricing, regions, and best-effort capacity headroom.

## Response `200`

OK

- DEListInferenceInstanceTypesResponse: Hardware instance types available for inference deployments.
  - `data` DEInferenceInstanceType[], required: Instance types available for inference.
    - `id` string, required: Stable hardware instance type identifier used by deployment configs.
    - `name` string, required: Human-readable instance type name.
    - `description` string, required: Human-readable summary of the hardware configuration.
    - `gpuType` string, required: GPU accelerator model, such as `H100` or `B200`.
    - `gpuCount` integer, required: Number of GPUs in one replica of this instance type.
    - `gpuMemoryGib` integer, required: Memory available on each GPU, in GiB.
    - `priceCentsPerHour` integer, required: On-demand price for one running replica, in US cents per hour.
    - `regions` DERegion[], required: Regions where this instance type is offered.
      - `name` string, required: Region name where an instance type is offered.
      - `headroom` DEHeadroom: Best-effort estimate of how many additional replicas currently fit in a region.
        - `value` integer: Capped count of replicas that currently fit.
        - `relation` 'RELATION_EQ' | 'RELATION_GTE', required: Whether the value is exact or a lower bound.
      - `compliance` DEComplianceAvailability[]: Compliance regimes certified in this region with best-effort headroom for each policy. Entries can overlap; read the entry matching the deployment policy instead of summing entries.
        - `policy` DECompliancePolicy, required: Compliance regimes required by a deployment placement policy.
          - `hipaa` boolean: Restrict placement to HIPAA-attested clusters.
        - `headroom` DEHeadroom, required: Best-effort estimate of how many additional replicas currently fit in a region.
          - `value` integer: Capped count of replicas that currently fit.
          - `relation` 'RELATION_EQ' | 'RELATION_GTE', required: Whether the value is exact or a lower bound.
  - `next_cursor` string: Cursor for the next page. Always null today because this catalog is returned in full.
  - `object` 'list', required: Object type. Always `list`.

## Other responses

- `default`: Default error response

## Changes

- **2026-10-22** `bb5257b3173a`: 1 info
  - endpoint added

[Operation history](https://skmtc.dev/together/apis/together-apis/changes/v2/public/inference-instance-types/get.md)

---

[API](https://skmtc.dev/together/apis/together-apis.md) - [All operations](https://skmtc.dev/together/apis/together-apis/llms.txt) - [OpenAPI document](https://skmtc.dev/together/apis/together-apis/revisions/93438ddac711?raw)