Client-side only
All tools work with static JavaScript and do not send workload assumptions to a backend.
Free browser-only tools
Estimate compute cost before you buy capacity.
These calculators run entirely in your browser. Use them to turn dated public GPU prices, your own quotes, and rough serving assumptions into comparable workload costs.
Key facts
Dated price data
Provider comparison rows are static examples, not live procurement quotes.
Workload-first
The calculators emphasize cost per useful unit instead of ranking vendors.
gpu cost calculator
GPU Cost Estimator
Estimate GPU rental cost from dollars per GPU-hour, hours, token volume, throughput, GPU count, and utilization.
- Runs entirely in the browser with no account, backend, or data collection.
- Supports time-based jobs and token-throughput jobs in the same form.
gpu cloud price comparison
GPU Price Compare
Compare dated public GPU-hour examples for H100, A100, L40S, RTX 4090, and live-market GPU listings across providers.
- Price examples are static snapshots checked 2026-07-06.
- Rows mix different markets, so the note column explains what is and is not included.
llm inference cost calculator
Inference Throughput Cost Calculator
Estimate rough self-hosted LLM inference cost per request from model size, context length, batch size, output tokens, and GPU hourly price.
- The formula is intentionally transparent and editable; it is not a vendor benchmark.
- Batch size improves throughput but can add latency, so the result should be compared with p95 targets.
How to use these tools together
Start with the GPU Price Compare table to choose candidate hourly rates. Use the GPU Cost Estimator for planned runtime, token volume, and daily burn. Use the Inference Throughput Cost Calculator when the workload is self-hosted LLM serving.
The output is a planning estimate. Before procurement, build a benchmark table with the live source quote, region, SKU, storage, egress, support terms, measured throughput, measured latency, utilization, and retry rate.