AI Glossary Term

What does “Usage-Based Billing” mean?

Definition

Pricing where costs scale with API calls, tokens, inference time, or consumed compute.

Browse the full AI Glossary Explore all free AI guides

How Usage-Based Billing works

Example

An AI provider charges per million input and output tokens. A chatbot that handles ten times more conversations in a month sees roughly ten times the usage portion of its bill.

Why it matters

Costs scale with usage, which can surprise teams. Longer prompts, outputs, and bigger models all raise the bill.

Common questions

How can I control costs?

Set budgets and alerts, shorten prompts, cache repeated work, and use smaller models where they suffice.

Usage-based vs subscription?

Usage-based bills by consumption; a subscription charges a fixed fee, often with limits.

Why are output tokens often pricier?

Generating each token takes more computation than reading input, so many providers price output higher.

Inference-Time ComputeThe amount of processing power consumed while producing each response.Test-Time ComputeAdditional inference computation used during response generation to improve quality or reasoning.Speculative DecodingAn inference acceleration method where a small draft model proposes tokens that a larger model verifies in parallel.Sliding Window AttentionAn attention pattern where each token attends only to a fixed-size window of nearby tokens to reduce compute.InferenceThe runtime phase where a trained model generates predictions or outputs.RAG (Retrieval-Augmented Generation)A method that retrieves external knowledge and feeds it into generation at inference time.

Learn more in our free guides

Entropy-Based Sampling Energy-Based Models Score-Based Generative Models WaveGlow Flow-Based Vocoder

Test yourself

Entropy-Based Sampling Quiz Energy-Based Models Quiz Score-Based Generative Models Quiz