---
title: "Current GPU rental quotes per card and provider (USD per GPU-hour)"
method: GET
path: "/api/gpu-prices.json"
---

# Current GPU rental quotes per card and provider (USD per GPU-hour)

`GET /api/gpu-prices.json`

## Response `200`

All tracked GPUs with per-provider rental quotes

- object
  - `updated` string, date-time — Last data refresh (ISO 8601)
  - `source` 'https://llmhosting.ai'
  - `attribution` string — License and upstream-source note
  - `gpus` object[]
    - `gpu` string — GPU name, e.g. "H100 SXM"
    - `vram_gb` number, nullable — VRAM in GB (null when unknown)
    - `quotes` object[] — Sorted cheapest-first
      - `provider` string — Provider slug, e.g. "vast", "runpod"
      - `price_hr` number — Cheapest observed USD per GPU-hour
      - `price_median_hr` number, nullable — Median USD per GPU-hour across live offers, null when single-quote
      - `price_type` string — Quote tier, e.g. "marketplace", "on-demand"
      - `offers` number, nullable — Live offer count behind the quote, null when not a marketplace

---

[API](https://skmtc.net/llmhosting/apis/llmhosting-ai-public-data-api.md) · [All operations](https://skmtc.net/llmhosting/apis/llmhosting-ai-public-data-api/llms.txt) · [OpenAPI document](https://skmtc-service-staging.skmtc.workers.dev/v1/apis/llmhosting/llmhosting-ai-public-data-api/revisions/0a23f7391161/schema)