GPU calculators that show their arithmetic.
Three calculators built on the same data as the rest of the site: model architectures read from their published configuration, and the live prices of 47 GPUs. Every formula is written under the result.
Pick the question you need answered.
LLM VRAM calculator
How much GPU memory a model needs for its weights, its KV cache at your context length and your number of concurrent requests — and the cheapest node that holds it.
64 models built inGPU rent vs buy
The monthly cost of owning a GPU (purchase, server, power, resale) against renting it by the hour, and the utilisation where buying starts to pay.
Live rental pricesJob cost calculator
What a job costs on a given GPU and node size, spot or on-demand, with persistent storage — and how much it saves against the cheapest price found elsewhere.
47 GPUsThe tables behind the calculators.
The calculators read the same tables you can browse: per-model sizing, per-GPU pricing against the market, and the published price index.
- Model GPU requirements
- 64 open modelsWeights by format, KV cache by context, every GPU that runs each one
- GPU catalogue
- 47 GPUsSpecifications, node sizes and prices against 102 providers
- Price guides
- 4 guidesH100 and B300 prices per hour, cheapest providers, H100 vs H200 vs B200
- catalog.json
- Machine-readableThe live catalogue with every price, for scripts and agents
Your first H100 for $0.55 an hour.
Prepaid in crypto, pay as you go: no card, no contract, no minimum commitment. Add credit, launch, stop whenever you want.
Billed per minute from the moment the instance is reachable. Minimum credit $30, no subscription.