{"data":{"slug":"tokens","etype":"concept","name":"Tokens & Tokenization","cat":"Core Technical","subtitle":"The unit AI is bought, sold and measured in.","summary":"Models read and write tokens — word fragments, roughly ¾ of an English word each. Context windows, API prices, throughput benchmarks and 'tokens per second' hardware claims all count them. When executives describe data centers as 'token factories,' this is the commodity they mean: the entire economics of inference reduces to the cost of producing a token.","attrs":null,"kf":null,"aliases":["tokenization","tokens per second"],"ticker":null,"exchange":null,"hq":null,"founded":null,"website":null,"status":null,"region":null,"asof":"2026-01-15","sources":[{"url":"https://platform.openai.com/tokenizer","label":"OpenAI tokenizer"}]},"meta":{"found":true,"count":1,"served_from":"site cache (no live upstream calls)","as_of":"2026-10-14T02:57:36+00:00","disclaimer":"Research and information only — not investment advice. Structured records carry basis labels (vendor spec / reported / estimate); market quotes are delayed/indicative.","attribution":"Records compiled from vendor documentation, official filings and reputable public reporting; market quotes via financial data providers. See /data-sources."}}