---
schema: 1
type: term
title: tokens per second
description: >-
  **Tokens per second** measures how fast an AI model generates text — how many
  tokens it outputs in one second. In the channel's captured material it appears
  as the headline figure for OpenAI's _Spark_ model, described as super-fast
  while running at **1,000 tokens per second**.
canonical: >-
  https://liquidgold.cc/channel/everyinc/explore/term/tokens-per-second-f57cc2b66a7d
markdown: >-
  https://liquidgold.cc/channel/everyinc/explore/term/tokens-per-second-f57cc2b66a7d.md
llms: null
parent: 'https://liquidgold.cc/channel/everyinc/llms.txt'
channel: everyinc
updated: '2026-10-23T22:12:35.346Z'
language: null
children: []
---


# tokens per second

Canonical: https://liquidgold.cc/channel/everyinc/explore/term/tokens-per-second-f57cc2b66a7d

Parent: [Channel index](https://liquidgold.cc/channel/everyinc/llms.txt)

## Definition

**Tokens per second** measures how fast an AI model generates text — how many tokens it outputs in one second. In the channel's captured material it appears as the headline figure for OpenAI's _Spark_ model, described as super-fast while running at **1,000 tokens per second**.

## Where the term shows up in this material