LLM token energy & cost estimator API

Estimate the electricity used to run N tokens (or queries) through an LLM, plus dollar cost and CO2, with fun comparisons (phone charges, LED-bulb hours, Google searches, EV miles). Rough Wh/token by model class x datacenter PUE x grid carbon x $/kWh, deterministic. Answers 'how much electricity does an LLM query use', 'cost to run a million tokens', 'carbon footprint of an AI prompt', 'energy per token'.

Price$0.01per request

MethodPOST

Route/v1/fun/llm-energy

StatusLive

MIME typeapplication/json

Rate limit120/minute

CacheNo cache

funenergyelectricityllmtokenscarboncostai

API URLhttps://x402.hexl.dev/v1/fun/llm-energy

Integration docs

Example request

{
  "tokens": 1000000,
  "model": "gpt-4o"
}

Example response

{
  "tokens": 1000000,
  "model": "gpt-4o",
  "whPerToken": 0.0003,
  "pue": 1.2,
  "energyWh": 360,
  "energyKwh": 0.36,
  "costUsd": 0.0576,
  "co2Grams": 144,
  "waterLiters": 0.648,
  "waterBottles500ml": 1.3,
  "comparisons": {
    "phoneChargesEquivalent": 18.95,
    "ledBulbHours": 36
  }
}

Input schema

{
  "type": "object",
  "properties": {
    "tokens": {
      "type": "number",
      "examples": [
        1000000
      ]
    },
    "queries": {
      "type": "number",
      "examples": [
        1000
      ]
    },
    "tokensPerQuery": {
      "type": "number",
      "examples": [
        500
      ]
    },
    "model": {
      "type": "string",
      "examples": [
        "gpt-4o",
        "llama-70b",
        "claude-opus"
      ]
    },
    "pricePerKwh": {
      "type": "number",
      "examples": [
        0.16
      ]
    },
    "gridGramsCo2PerKwh": {
      "type": "number",
      "examples": [
        400
      ]
    },
    "pue": {
      "type": "number",
      "examples": [
        1.2
      ]
    }
  }
}

Output schema

{
  "type": "object",
  "additionalProperties": true
}