This folder contains the comparative sustainability analysis of several
large language models (LLMs) used in commercial applications.
It is part of the ELO2 Green AI project, focusing on estimating
the energy, carbon, and water footprints of each model.
- models.md – Main document providing technical summaries and
sustainability estimates for:
- GPT-4 (OpenAI)
- Claude 3 Haiku (Anthropic)
- Gemini Nano (Google)
This documentation:
- Highlights how different LLM architectures and deployments affect energy and water use.
- Demonstrates how model size and hosting influence environmental impact.
- Supports ongoing evaluation of Green AI strategies for efficient computing.