P1XXL fka. chrissx

LLM inference cost calculator

Compare cloud API pricing against running locally (electricity cost), and find the breakeven price if you were to resell inference. Local cost assumes the machine runs 24/7 and serves your monthly token volume; all other time is counted at idle draw.

Cloud pricing ($/M tokens)
Monthly usage (M tokens)
Local hardware

Cloud (monthly)

Local (monthly)

At load

input + output time, per month

Total energy

per month

Input power

kWh / M tokens

Output power

kWh / M tokens

Breakeven price

blended, per M tokens, to cover energy cost

Input

/M tokens

Cache

97% off input

Output

/M tokens