GPU Economy

ClaudeCommunityby GPU Economy

GPU Economy reads the public price pages and APIs of 60+ GPU clouds every hour — hyperscalers (AWS, Azure, Google Cloud, Oracle), neoclouds (Lambda, CoreWeave, Nebius, Crusoe, RunPod and others) and marketplaces (Vast.ai, Clore.ai) — and normalizes every listed rate to US dollars per GPU per hour. This server gives an AI agent those prices directly, instead of scraping pages. Tools (all read-only; no key, no account): - list_gpus — every tracked GPU model (H100, H200, B200, B300, A100, L40S, RTX 4090/5090, MI300X and more) with VRAM, how many providers list it and the median on-demand rate. - gpu_prices — current rates for one GPU across providers, cheapest first; on-demand by default, or spot, community, serverless and reserved. - provider_prices — every current rate at one provider. - price_index — the weekly Cloud GPU Price Index: the median on-demand H100 SXM rate, this week and the last eight. - recent_price_changes — quotes whose listed price changed in the last 30 days, with the figure before and now. Every price comes with the time it was read and a link to the GPU Economy page that shows it beside its source. Prices not confirmed in the last 48 hours are left out. Rates are listed prices, not availability, and each row says what it covers: a whole instance, the GPU only, a marketplace floor or a "from" price. Listings in other currencies keep the listed figure and the official exchange rate used. No provider pays to be listed or ranked. Data is free under CC BY 4.0: https://gpueconomy.com/data