Canon term · Context & memory

Context Window

The bounded span of tokens a model can attend to in a single inference, containing the prompt, retrieved material, and conversation so far.

  • StatusIndustry term
  • AdoptionWidely used
  • LayerIntelligence

Why it matters

The context window is a hard architectural limit, not a memory system. Confusing the two produces designs that silently drop the evidence a decision was based on.

Where it comes from

Language model architecture; universal usage.

No published instrument defines this term yet; the note above is its origin.

What it is not

These are routinely confused with Context Window. The distinctions have consequences for how a system is governed.

  • Agent Memory

    The window is per-inference and ephemeral. Memory persists across invocations and is a separate, governable artifact.

  • Training Data

    The window is runtime input. Training data is baked into weights and cannot be inspected per-decision.

How it connects

Not to be confused with

Cite and reuse

Cite the identifier QIS-TERM-00009, not the URL: identifiers are stable, URLs may change. This entry is free to read, quote and index under the dual license. Corrections sent to editor@qisfund.com are published.

Technical details: identifier, graph relationships and record
QIS-TERM-00009: outbound relationships as published in api/graph.json and api/lexicon.json
PredicateTarget identifierTargetPredicate meaning
notEquivalentToterm:agent-memoryAgent MemoryThe two are routinely conflated and are distinct.
Canonical identifier
QIS-TERM-00009
Tier
Canonical industry term
Domain · layer
Context & memory · Intelligence
Semantic aliases
None recorded
First published
2026-08-02
Last reviewed
2026-08-02 · 180-day cycle