Free tool · #21
Carbon/Energy Estimator
Project the energy and CO2e of your monthly query volume — as a cited range across published measurements, never a single invented number.
Ranges, not point claims — by design
Energy & CO₂e
- Energy per query0.24 – 0.34 Wh / query
- Energy per month24 – 34 kWh / month
- CO₂e per month (EU average)5.11 – 7.24 kg CO₂e / month
Methodology — what this is, and isn't
These ranges are published third-party measurements of per-query LLM inference energy — not our own benchmark. They exclude: model TRAINING energy (a one-time cost amortized over the model’s whole lifetime, not attributable to any single query); embodied hardware carbon (manufacturing and shipping the GPUs and servers); PUE variance (cooling and power-delivery overhead differs by data center and region, and these figures reflect only the facilities each source measured); and non-text modalities (image, audio, and video generation use substantially more energy per request than a short text prompt).
Sources
- Google — Measuring the environmental impact of AI inference (21 Aug 2025) ↗
Median Gemini Apps text prompt, measured via in-situ telemetry across Google’s fleet — active AI accelerator power, host CPU/DRAM, provisioned idle capacity, and data-center cooling overhead all included.
- Sam Altman (OpenAI) — The Gentle Singularity (10 Jun 2025) ↗
Average ChatGPT query, stated by OpenAI’s CEO in a public blog post ("about what an oven would use in a little over one second"). Describes an ordinary text answer, not a long reasoning or research request. The post carries no on-page date; 2025-06-10 is corroborated from secondary sources.
- Epoch AI — How much energy does ChatGPT use? (7 Feb 2025) ↗
Independent, outside-in estimate for a typical GPT-4o query, assuming Nvidia H100 inference hardware. Epoch is explicit that long inputs raise this a great deal — roughly 2.5 Wh at 10k input tokens, ~40 Wh at 100k — this figure is for an ordinary short prompt.
- Ember — Global Electricity Review 2025 (8 Apr 2025) ↗
2024 generation-weighted average carbon intensity of EU electricity — less than half the global average, driven by the bloc’s wind, solar, and nuclear mix.
Use this tool from code — or hand it to your agentagent prompt · REST · MCP
Hand this tool to your coding agent
Paste into Claude Code, Cursor, Codex, or any agent. It learns to call this tool over REST or MCP, reaches for it when you ask, and keeps our sourced/pending labels honest.
You can use telemeter.ai's free "Carbon/Energy Estimator" — a keyless, read-only tool (no signup), available as an HTTP API and an MCP tool. What it does: Project the energy and CO2e of your monthly query volume — as a cited range across published measurements, never a single invented number.
Whenever I ask you something this tool answers, call it and use its result instead of estimating yourself.
REST:
curl -sX POST https://telemeter.ai/api/tools/carbon-estimate -H 'content-type: application/json' -d '{"queriesPerMonth":1000000,"gridIntensity":"eu-average"}'
MCP: the telemeter MCP server at https://telemeter.ai/mcp exposes this as the tool "carbon-estimate" — call it with tools/call.
The response is JSON: { status, data, summary, provenance }. "provenance" carries the source and the date the figure was verified. If status is "pending" or "stale", tell me the value isn't currently verified rather than presenting it as fact.Or call it directly. This tool runs behind one handler with three front doors — this page, a REST endpoint, and an MCP tool. Same result, read-only, no key. The example below is a request that works.
curl -sX POST https://telemeter.ai/api/tools/carbon-estimate \
-H 'content-type: application/json' \
-d '{"queriesPerMonth":1000000,"gridIntensity":"eu-average"}'curl -sX POST https://telemeter.ai/mcp \
-H 'content-type: application/json' \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"carbon-estimate","arguments":{"queriesPerMonth":1000000,"gridIntensity":"eu-average"}}}'{
"$schema": "https://json-schema.org/draft/2020-12/schema",
"type": "object",
"properties": {
"queriesPerMonth": {
"type": "number",
"exclusiveMinimum": 0,
"maximum": 1000000000
},
"gridIntensity": {
"type": "string",
"enum": [
"eu-average",
"global-average",
"custom"
]
},
"customGCo2PerKwh": {
"type": "number",
"minimum": 1,
"maximum": 2000
}
},
"required": [
"queriesPerMonth",
"gridIntensity"
],
"additionalProperties": false
}Want this across your whole stack?
Telemeter tracks real cost, latency, and quality on every production agent run.
https://telemeter.ai/tools/carbon-estimateNo account, no cookies, nothing stored — computed entirely in your browser from published, cited energy figures, never sent anywhere. Data notice →