Glossary

AI & LLM Glossary

Essential terminology for understanding AI costs, LLM APIs, and optimization. Clear definitions designed for developers and engineering teams.

13
Terms
11
Categories
2026
Updated
ACFILMOPRST
A

AI FinOps

AI FinOps is the practice of managing and optimizing AI infrastructure costs, particularly LLM API spending.

C

Context Window

The context window is the maximum number of tokens a model can process in a single request, including both input and output.

F

Fine-tuning

Fine-tuning customizes a base model on your specific data to improve performance for your use case.

I

Input Tokens

Input tokens are the tokens in your prompt sent to the model. They're typically cheaper than output tokens.

L

LLM Observability

LLM observability is the practice of monitoring, tracking, and analyzing LLM usage including costs, latency, errors, and quality.

Latency

Latency is the time between sending a request and receiving the complete response. It includes time-to-first-token and generation time.

M

Model Routing

Model routing automatically selects the best model for each request based on complexity, cost, or other criteria.

Multimodal

Multimodal models process multiple types of content: text, images, audio, and video in a single request.

O

Output Tokens

Output tokens are the tokens generated by the model in its response. They're typically 3-10x more expensive than input tokens.

P

Prompt Caching

Prompt caching stores processed prompts to reduce cost and latency for repeated requests with similar prefixes.

R

Reasoning Model

Reasoning models (like o3, DeepSeek R1) use chain-of-thought processing internally to solve complex problems.

S

Streaming

Streaming delivers model responses token-by-token as they're generated, improving perceived latency for end users.

T

Token

A token is the basic unit of text that LLMs process. In English, one token is roughly 4 characters or 0.75 words.

Put This Knowledge to Work

Understanding LLM terminology is the first step. Burnwise helps you track and optimize every metric that matters.

Start Free Trial