← sources

How much does your local LLM server really cost? (power draw + tok/s, self-hosted vs cloud API, free in-browser) :)

source: reddit (reddit:LocalLLaMA)published 2026-10-09 08:39ingested 2026-10-10 02:38

retained text

I've been tuning local LLM setup for many months, and the number people actually want is the electricity bill. This calculator takes your GPU's power draw (W) and token speed (tok/s for prompt and decode), plugs in your energy price per kWh, and shows cost per hour, per request, monthly, and cost per 1M input tokens — with and without KV cache.

  
There are 3 GPU presets (RTX 4070 Ti Super, RTX 5090, RTX 5090 eco) if you want to start from realistic numbers and adjust from there. The math treats cached tokens at 1/100 the time of uncached, so the realistic scenario is fixed at 60k input / 50k cached / 2k output / 90% uptime.

  
The cloud API comparison compares each self-hosted profile against a reference API ($0.25/1M input, $1.20/1M output) at the same scenario, so you can see whether self-hosting actually saves money or costs more.

  
Everything runs in your browser, nothing is uploaded and there is no sign-up. From my experience, the per-request number is the one worth comparing with the cloud — the monthly bill is just the per-request cost times your actual request rate.

  
[https://appdoesit.com/apps/llm-cost-calculator](https://appdoesit.com/apps/llm-cost-calculator) — it's one of the 139 free tools in the catalog.

  
Try it by yourself :) — how does your server compare?

document fetched 2026-10-10 00:41 · canonical https://reddit.com/r/LocalLLaMA/comments/1x1g2kd/how_much_does_your_local_llm_server_really_cost