Merge pull request #14 from MIT-Emerging-Talent/Commercial-models

Milestone 2: Commercial Models Research — Energy, Carbon, and Accuracy Analysis #8

Author: doctorbanu

commercial_models/README.md

@@ -0,0 +1,29 @@
+# Commercial Models – Green AI Analysis
+
+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.
+
+---
+
+## 📄 Contents
+
+- **models.md** – Main document providing technical summaries and
+ sustainability estimates for:
+  - GPT-4 (OpenAI)
+  - Claude 3 Haiku (Anthropic)
+  - Gemini Nano (Google)
+  
+---
+
+## 🎯 Purpose
+
+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.
+
+---

commercial_models/models.md

@@ -0,0 +1,197 @@
+<!-- markdownlint-disable MD013 MD024 MD025 MD026 MD041 MD001 -->
+<!-- This disables line length, duplicate headings, multiple top-level headings, heading punctuation, first heading, and heading increment rules -->
+<!-- markdownlint-disable MD013 -->
+<!-- markdownlint-disable MD025 -->
+<!-- Multiple top-level headings needed for each model section -->
+
+# Comparative Environmental and Technical Overview of Modern AI Models
+
+This document presents verified technical and environmental data for three leading AI models — **OpenAI GPT-4**, **Anthropic Claude 3 Haiku**, and **Google Gemini Nano** — focusing on energy, water, and sustainability context.
+
+---
+
+## Model 1: GPT-4 (OpenAI)
+
+### Model Name & Provider
+
+**GPT-4**, developed by **OpenAI**.
+
+### Hosting / Deployment
+
+Hosted via **Microsoft Azure OpenAI Service**, which operates on Azure’s global data centers (specific regions not publicly disclosed).  
+Source: [Introducing GPT-4 in Azure OpenAI Service – Microsoft Azure Blog](https://azure.microsoft.com/en-us/blog/introducing-gpt-4-in-azure-openai-service/)
+
+### Estimated Model Size / Architecture
+
+GPT-4 is widely considered a **frontier model** employing a **Sparse Mixture-of-Experts (MoE)** architecture.  
+This structure activates only a subset of parameters per inference, optimizing efficiency while maintaining scale.  
+Estimated total parameters exceed **1 trillion**.  
+
+Sources:  
+
+- [*Optimal Sparsity of Mixture-of-Experts Language Models for Reasoning Tasks* – arXiv (2508.18672v2)](https://arxiv.org/html/2508.18672v2)  
+- [*Scaling Laws for Optimal Sparsity for Mixture-of-Experts Language Models* – arXiv (2501.12370v2)](https://arxiv.org/html/2501.12370v2)
+
+### Estimated Energy (Inference)
+
+Estimates vary between studies:  
+
+- **Epoch AI (2024):** ≈ 0.3 Wh (0.0003 kWh) per query under typical load.  
+  Source: [Epoch AI – *How Much Energy Does ChatGPT Use?*](https://epoch.ai/gradient-updates/how-much-energy-does-chatgpt-use)  
+- **Other analysts:** 0.3 – 1.8 Wh (0.0003 – 0.0018 kWh), depending on token count and hardware.
+
+**Note:** OpenAI has not released official inference energy data.  
+Values depend on hardware (GPU vs TPU), data-center PUE, and carbon intensity.
+
+### Training Energy Estimates
+
+Extrapolated from compute budgets and model size:  
+≈ 51,772,500 – 62,318,750 kWh (≈ 51.8 – 62.3 GWh) for full-scale training.  
+Source: [*The Carbon Footprint of ChatGPT* – Sustainability by Numbers](https://sustainabilitybynumbers.com/how-much-energy-does-chatgpt-use/)
+
+### Water Usage
+
+Water consumption derives from data-center cooling processes:  
+
+- A single ChatGPT query may indirectly consume ≈ 0.5 L of water.  
+- Generating a 100-word email ≈ 0.14 kWh energy + 0.52 L water.  
+Source: [The Verge – *Sam Altman on ChatGPT Energy and Water Use*](https://www.theverge.com/news/685045/sam-altman-average-chatgpt-energy-water?)
+
+### PUE / CI Context Used in Studies
+
+Environmental analyses generally apply:  
+
+- **PUE (Power Usage Effectiveness):** ≈ 1.1 – 1.3 (Azure hyperscale data centers)  
+- **CI (Carbon Intensity):** ≈ 0.3 – 0.4 kg CO₂e / kWh (depending on regional grid mix)
+
+---
+
+## Model 2: Claude 3 Haiku (Anthropic)
+
+### Model Description
+
+**Claude 3 Haiku** is the smallest and fastest member of Anthropic’s Claude 3 model family ( Haiku, Sonnet, Opus ), released March 2024.  
+It is optimized for low-latency, energy-efficient applications such as chatbots, summarization, and enterprise automation.  
+
+Sources:  
+
+- [Anthropic Blog – *Introducing the Claude 3 Model Family*](https://www.anthropic.com/news/claude-3-family)  
+- [Anthropic 3 Announcement](https://www.anthropic.com/news/claude-3-7-sonnet)
+
+### Hosting / Deployment
+
+Claude 3 Haiku is available through Anthropic’s API and **AWS Bedrock**.  
+AWS data centers maintain an average PUE of ≈ 1.2.  
+Sources:  
+
+- [AWS Bedrock – *Use Claude on Bedrock*](https://aws.amazon.com/bedrock/claude/)  
+- [AWS Sustainability Report 2024](https://sustainability.aboutamazon.com/reporting)
+
+### Estimated Model Size / Architecture
+
+Community estimates place Claude 3 Haiku at ≈ 20 B parameters.  
+The largest model in the family (Claude 3 Opus) is ≈ 2 T parameters.  
+Source: [ClaudeAI Community Discussion (Reddit)](https://www.reddit.com/r/ClaudeAI/)
+
+### Estimated Energy (Inference)
+
+Anthropic does not publish per-query energy figures.  
+Based on 10–30 B parameter transformers: ≈ 0.05 – 0.1 Wh (0.00005 – 0.0001 kWh) per query.  
+Haiku is ≈ 5× more efficient than Claude 3 Sonnet or Opus.  
+Sources:  
+
+- [Epoch AI – Machine Learning Trends (for compute/power scaling)](https://epoch.ai/trends)  
+- [Anthropic 3 Announcement](https://www.anthropic.com/news/claude-3-7-sonnet)
+
+### Training Energy Estimates
+
+Claude 3 models are trained on GPU clusters (NVIDIA A100/H100) via AWS.  
+Typical training energy for models of this scale: ≈ 3,000 – 10,000 MWh.  
+Sources:  
+
+- [Epoch AI – Machine Learning Trends (for compute/power scaling)](https://epoch.ai/trends)  
+- [Anthropic Responsible Scaling Policy](https://www-cdn.anthropic.com/872c653b2d0501d6ab44cf87f43e1dc4853e4d37.pdf)
+
+### Water Usage
+
+Anthropic does not publish direct figures; relies on AWS cooling efficiency and water recycling policies.  
+Sources:  
+
+- [AWS Water Stewardship Report](https://sustainability.aboutamazon.com/2024-amazon-sustainability-report-aws-summary.pdf)  
+- [Anthropic Responsible Scaling Policy](https://www-cdn.anthropic.com/872c653b2d0501d6ab44cf87f43e1dc4853e4d37.pdf)
+
+### PUE / CI Context Used in Studies
+
+- **PUE:** ≈ 1.2 (AWS average)  
+- **CI:** ≈ 0 – 0.2 kg CO₂e / kWh (based on regional renewable mix)  
+AWS targets **100 % renewable energy by 2025**.  
+Sources:  
+- [AWS Global Infrastructure Efficiency Data](https://sustainability.aboutamazon.com/2024-amazon-sustainability-report-aws-summary.pdf)  
+- [Anthropic Responsible Scaling Policy](https://www-cdn.anthropic.com/872c653b2d0501d6ab44cf87f43e1dc4853e4d37.pdf)
+
+---
+
+## Model 3: Gemini Nano (Google DeepMind)
+
+### Model Name & Provider
+
+**Gemini Nano**, developed by **Google DeepMind**, is the smallest member of the Gemini family (Nano, Pro, Ultra).  
+Sources:  
+
+- [Google AI Blog – *Introducing Gemini*](https://blog.google/technology/ai/google-gemini-ai/)  
+
+### Hosting / Deployment
+
+Runs **on-device** through Android’s **AICore** system (launched in Android 14).  
+Deployed on optimized hardware (e.g., Pixel 8 Pro, Pixel 9 Series).  
+This local processing approach eliminates cloud compute energy and network latency.  
+Additional coverage: - [Android Developers – *Gemini Nano Overview*](https://developer.android.com/ai/gemini-nano)
+
+### Estimated Model Size / Architecture
+
+Deployed in quantized versions:  
+
+- **Nano-1:** ≈ 1.8 B parameters  
+- **Nano-2:** ≈ 3.25 B parameters  
+Reference: [Exploding Topics – AI Model Parameters Database](https://explodingtopics.com/blog/gpt-parameters)
+
+### Estimated Energy (Inference)
+
+- **Median Cloud Gemini Inference:** ≈ 0.24 Wh per text prompt.  
+- **On-Device Nano Estimate:** ≈ 0.01 Wh per query (benchmarks + design targets).  
+Note: Official Nano inference measurements are not yet public.  
+Source: [Google Cloud Blog – *Measuring the Environmental Impact of AI Inference*](https://cloud.google.com/blog/products/infrastructure/measuring-the-environmental-impact-of-ai-inference/)
+
+### Training Energy Estimates
+
+Gemini Nano was distilled from larger Gemini models trained on **Google TPU v5e clusters**.  
+Training energy estimated ≈ 200 – 1,200 MWh (total, amortized across billions of devices).  
+Sources:  
+
+- [Google Cloud TPU Documentation](https://cloud.google.com/tpu/docs/)  
+- [Google Cloud Blog – Environmental Impact of AI Inference](https://cloud.google.com/blog/products/infrastructure/measuring-the-environmental-impact-of-ai-inference/)
+
+### Water Usage
+
+- **Inference:** Zero data-center water use (on-device).  
+- **Training:** Uses Google data centers with average WUE ≈ 0.26 mL per median cloud query.  
+Source: [Google Cloud Blog – *Measuring the Environmental Impact of AI Inference*](https://cloud.google.com/blog/products/infrastructure/measuring-the-environmental-impact-of-ai-inference/)
+
+### PUE / CI Context Used in Studies
+
+- **Average PUE:** 1.10 – 1.12 (Google Data Centers)  
+- **Carbon Intensity (CI):** ≈ 0.03 g CO₂e / query (market-based)  
+Reflects Google’s near-total renewable energy purchasing.  
+Source: [Google Cloud Blog – Environmental Impact of AI Inference](https://cloud.google.com/blog/products/infrastructure/measuring-the-environmental-impact-of-ai-inference/)
+
+---
+
+### Summary
+
+| Model | Developer | Hosting Type | Est. Parameters | Inference Energy (Wh/query) | Training Energy (MWh) | PUE | CI (kg CO₂e/kWh) |
+|:------|:-----------|:--------------|:----------------|:-----------------------------|:----------------------|:----|:----------------:|
+| GPT-4 | OpenAI | Cloud (Azure) | ≈ 1 T + (MoE) | 0.3 – 1.8 | ≈ 51 K – 62 K MWh | 1.1–1.3 | 0.3–0.4 |
+| Claude 3 Haiku | Anthropic | Cloud (AWS Bedrock) | ≈ 20 B | 0.05 – 0.1 | 3 K – 10 K | ≈ 1.2 | 0–0.2 |
+| Gemini Nano | Google DeepMind | On-Device (Android AICore) | 1.8–3.25 B | ≈ 0.01 (on-device) | 200–1,200 | 1.10–1.12 | ≈ 0.03 g CO₂e /query |
+
+---