The SIMPLE Law for Energy Cost of LLM Inference
cs.AI, cs.LG, cs.PF
Submitted: 2026-11-02
Updated: 2026-11-16
Terminology
Sources
- The Falcon Series of Open Language Models
- The Price Reversal Phenomenon: When Cheaper Reasoning Models Cost More
- DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model
- DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning
- Small Talk, Big Impact: The Energy Cost of Thanking AI
- Measuring the environmental impact of delivering AI at Google Scale
- The Llama 3 Herd of Models
- How Hungry is AI? Benchmarking Energy, Water, and Carbon Footprint of LLM Inference
- Scaling Laws for Neural Language Models
- Agentic Coding in the Wild: Characterizing GitHub Copilot Traces at Production Scale
- Granite Code Models: A Family of Open Foundation Models for Code Intelligence
- The Token-Energy Law of LLM Decode: Active Bytes and Realized Time Account for Joules per Token
- EnergyLens: Interpretable Closed-Form Energy Models for Multimodal LLM Inference Serving
- Intelligence per Watt: Measuring Intelligence Efficiency of Local AI
- Fast Transformer Decoding: One Write-Head is All You Need
- SwiReasoning: Switch-Thinking in Latent and Explicit for Pareto-Superior Reasoning LLMs
- Towards Greener LLMs: Bringing Energy-Efficiency to the Forefront of LLM Inference
- Gemma: Open Models Based on Gemini Research and Technology
- Qwen3 Technical Report
- From Tokens to Watt-hours: Analytical Energy Estimation for LLM Inference on Modern GPUs
Related papers
- Multi-Objective Structured Pruning of LLMs for Latency and Model Size Optimization
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Embeddings for Preferences, Not Semantics
- Can We Trust Item Response Theory for AI Evaluation?
- Diagnosing Task Insensitivity in Language Agents