keepUpAI
CoursesLeaderboardGuidesInvitePricing
Start freeYou
HomeGlossaryLatency (TTFT)

AI glossary

What is Latency?

How long an AI request takes. Time to first token (TTFT) measures when output starts, and total time includes generating every token.

Output generation usually dominates total time. Track p95 latency, not just averages, and stream responses to improve perceived speed.

▶Practice it hands-onObservability: Traces, Latency & Cost

Related terms

StreamingPrompt Caching

← All 43 AI terms

keepUpAI

Hands-on AI engineering practice for students, developers and career-changers. Learn by doing, one fun step at a time.

Courses

  • All lessons
  • Foundations
  • Prompting & LLM APIs
  • Retrieval (RAG)
  • Agents
  • Evals & Production
  • ML for AI Engineers
  • Build an AI SaaS

Guides

  • All guides
  • How to Learn AI Engineering in 2026
  • What Is an AI Engineer
  • What Is RAG (Retrieval-Augmented Generation)
  • Prompt Engineering Guide
  • What Are AI Agents
  • What Is MCP (Model Context Protocol)

Glossary

  • All AI terms
  • RAG
  • Embedding
  • AI Agent
  • Prompt Injection
  • Temperature
  • MCP

keepUp AI

  • Leaderboard
  • Free tools
  • LLM cost calculator
  • Token counter
  • AI Engineer Toolkit
  • AI vs ML vs SWE
  • Pricing
  • Invite friends
© 2026 keepUp AI · Learn AI engineering by doing · Privacy · Terms