Token usage
Also written: token count
Units of text or data processed by AI, impacting cost and speed.
Tokens are the basic units of text that AI models like ChatGPT process. Think of them as pieces of words or punctuation. For example, the word 'tokenization' might be broken into 'token', 'ization'.
AI models read and write by processing these tokens. The more tokens an AI processes, the longer it takes and the more it costs. This is why tracking token usage is important for managing AI expenses and performance.
Tokens represent discrete units of text or data, typically sub-word units derived from a vocabulary. Models process input prompts and generate output sequences as a series of tokens. Token count is a primary metric for computational cost, influencing API pricing and inference latency. Different tokenization strategies (e.g., Byte Pair Encoding, WordPiece) result in varying token lengths for the same text, impacting efficiency and model behavior.
Token usage is a core metric for understanding and managing the cost and performance of large language models. As companies like OpenAI and Google release new models and features, tracking token counts helps users and developers optimize for efficiency and budget, especially when integrating AI into business applications.
- How to connect AI usage to business valueOpenAI · OpenAI News ·
- GPT-Realtime-2.1 ModelOpenAI · Model Documentation ·
- EvoLib: Turning experience into evolving knowledgeMicrosoft · Microsoft Research Blog ·
- Ep 785: What’s new in Gemini 3.5 Flash, Google Omni and Antigravity 2.0: Hands On With the latest from Google I/OGoogle · Everyday AI ·
- Google Gemini
- Google's Gemini models process information in tokens, and understanding token usage is key to managing the performance and cost of their AI services, as highlighted in discussions around new Gemini versions.
- Microsoft Copilot
- Microsoft Copilot, like other AI systems, relies on token processing for its operations. Tracking token usage is fundamental to managing the computational resources and associated costs of its AI functionalities.
- OpenAI ChatGPT
- OpenAI's models, including GPT, process text as tokens. Token usage is a primary factor in how OpenAI bills for API access and is a critical metric for users looking to control costs and understand model performance.
- Word
- A word is a linguistic unit, while tokens can be parts of words, whole wor
- Character
- A character is a single letter, number, or symbol, whereas tokens are
- Byte
- A byte is a unit of digital information, and tokens are a higher-level abs
The concept of tokenization in natural language processing dates back to early NLP research. Modern LLMs adopted and refined tokenization as a fundamental step for processing text data. Source ↗
Use ChatGPT or another AI chatbot. Type a sentence and observe how quickly it responds. Then, try typing a very long paragraph. Notice any difference in speed? This difference is partly due to the increased number of tokens the AI had to process.
The Playbook: reusable how-tos, prompts, configs and tips pulled out of the news.
- PromptRole-play a student struggling with math for AI tutor simulationSimulate a student's math struggles and uncertainty for AI tutor evaluation. · 5 min
- TemplateReusable workflow templates for healthcare teamsTeams can create and share templates for common workflows like documentation or prior authorizations. · 15 min
- TemplateTemplate for building an AI tool tailored to your workA structured approach to defining and building a custom AI tool for specific business needs. · 15 min
- TemplateEvals CSV prompt list for agent skillsA CSV file template for listing prompts used in evaluating agent skills. · 20 min
- HeuristicDouble coding agent task completion with docs linksProviding coding agents with a documentation link significantly increases their task completion rate. · 2 min
Questions people ask
- How many tokens is a word?
- It varies. On average, in English, 100 tokens are about 75 words. Some words might be one token, while others are broken into multiple tokens.
- Why does token usage matter?
- Token usage directly affects the cost of using AI models and the speed at which they can process information and generate responses.
First seen on Subvolts Jul 30, 2026. Updated Sep 23, 2026. Written by an AI model from the items above and reviewed against them weekly; report an error.