AI Provider Profile

OpenAI Sustainability Profile

Low Transparency

Environmental impact data for 30 tracked OpenAI models.

30
Models tracked
41.3 Wh
Avg. energy / query
5
Categories
3
Open-source

Sustainability Profile

Renewable Energy

Not published

Emissions Trend

Not published

Water Trend

Not published

Transparency Rating

Low

No dedicated sustainability report as of 2026.

How OpenAI Compares

OpenAI's average metrics compared to the full dataset of 152 models

Energy / query

This provider41.3 Wh
Dataset avg.48.1 Wh
Difference-14%

CO2 / query

Provider avg.18.2 g
Dataset avg.14.5 g
Difference+26%

Water / query

Provider avg.152 mL
Dataset avg.105 mL
Difference+44%

Environmental Overview

OpenAI maintains one of the largest AI portfolios in our dataset with 30 tracked models spanning 5 categories. On average, their models consume 41.3 Wh per query, which sits below the dataset-wide average of 48.1 Wh. The efficiency range is vast: GPT-4o Mini at 0.14 Wh versus Sora 2 at 600.0 Wh, a 4286x spread. 3 of their 30 models are open-source, which supports transparency and community-led efficiency improvements. Environmental transparency remains limited for OpenAI. No dedicated sustainability report as of 2026.

All OpenAI Models

Text / Chat21

ModelEnergyCO2WaterConfidence
GPT-4o Mini0.14 Wh0.06 g0.50 mLestimatedGPT-4.1 Mini0.18 Wh0.08 g0.65 mLestimatedGPT-4.1 Nano0.27 Wh0.12 g2 mLpeer-reviewedGPT-5.4 Nano0.30 Wh0.13 g1 mLestimatedGPT-OSS 20B0.35 Wh0.15 g1 mLestimatedGPT-4o0.43 Wh0.19 g2 mLpeer-reviewedGPT-4.10.45 Wh0.20 g2 mLestimatedGPT-4 Turbo0.50 Wh0.22 g2 mLestimatedGPT-5 Mini0.80 Wh0.35 g3 mLestimatedGPT-5.4 Mini1.0 Wh0.44 g4 mLestimatedo4-mini1.5 Wh0.65 g6 mLestimatedGPT-OSS 120B1.8 Wh0.79 g7 mLestimatedo3-mini2.0 Wh0.90 g8 mLestimatedo1-mini3.0 Wh1.3 g12 mLestimatedGPT-53.5 Wh1.5 g13 mLestimatedGPT-5.25.0 Wh2.2 g18 mLestimatedGPT-5.45.5 Wh2.4 g20 mLestimatedo115.0 Wh6.5 g55 mLestimatedGPT-4.520.5 Wh8.0 g80 mLpeer-reviewedo3 (Reasoning)21.4 Wh10.0 g100 mLpeer-reviewedo3-pro35.0 Wh15.3 g130 mLestimated

Code Assistant2

ModelEnergyCO2WaterConfidence
GPT-5.2 Codex4.0 Wh1.7 g15 mLestimatedGPT-5.3 Codex4.5 Wh2.0 g16 mLestimated

Image Generation3

ModelEnergyCO2WaterConfidence
DALL-E 33.0 Wh1.2 g15 mLestimatedGPT Image 1.53.8 Wh1.7 g14 mLestimatedGPT-4o Image Generation4.5 Wh2.0 g16 mLestimated

Video Generation2

ModelEnergyCO2WaterConfidence
Sora500.0 Wh225.0 g1.85 LestimatedSora 2600.0 Wh262.0 g2.16 Lestimated

Audio2

ModelEnergyCO2WaterConfidence
OpenAI TTS0.30 Wh0.13 g1 mLestimatedWhisper0.50 Wh0.22 g2 mLestimated

Frequently Asked Questions

How many AI models does OpenAI have?

OpenAI has 30 AI models tracked in the Know Your Compute database, spanning 5 categories: Text / Chat, Image Generation, Video Generation, Audio, Code Assistant.

What is OpenAI's most energy-efficient model?

OpenAI's most energy-efficient model is GPT-4o Mini, consuming 0.14 Wh per query. Their least efficient model is Sora 2 at 600.0 Wh per query.

Does OpenAI use renewable energy?

OpenAI's renewable energy claim: Not published. No dedicated sustainability report as of 2026.

Calculate your OpenAI footprint

Use our calculator to estimate your personal environmental impact from OpenAI models.

Calculate My Compute