Skip to content
ReviewRamble

The weekly brief

AI, reviewed, in your inbox.

  • New reviews of tools, models, MCP servers and repos, with the verdict up front
  • Re-checked prices and ratings when something changes
  • New analysis pieces as they are published

Models, agents, MCP servers and the tools actually worth your time.

One email a week. Unsubscribe with the link in any brief. How we use your email

Context rot

Context rot is the drop in a language model's accuracy and reliability as its input grows longer, even when the input still fits inside the context window and the needed information is present.

By Divya Sharma · Checked Nov 15, 2026 · How we review

A large context window tells you how much text a model will accept, not how well it will use it. Context rot is the gap between the two. As prompts get longer, models miss facts, mix up similar passages and follow instructions less closely.

The evidence

Chroma's July 2025 report tested 18 models, including GPT-4.1, Claude 4, Gemini 2.5 and Qwen3, and found that performance fell as input length grew in every experiment, often in uneven and surprising ways. Simple needle-in-a-haystack tests, where a model only has to find one sentence, hide much of this, because real tasks need the model to reason over what it finds and to ignore look-alike distractors.

Why it matters for agents

Agents pile up tool output, file contents and old turns, so their context grows on its own. Past a point, more context makes the agent worse. This is a main reason for context engineering: put only what the next step needs into the window.

How to reduce it

Retrieve the few relevant passages with retrieval-augmented generation instead of pasting whole documents. Summarise long histories with context compaction. Push side tasks to a subagent so their output never reaches the main thread. Put the most important instructions where the model sees them clearly, and test your own task at the context sizes you actually use, since the decline varies by model and task.

Related terms

Context compaction
Context compaction is the step where an agent summarises a conversation that is nearing its context window limit and continues in a fresh window that starts from that summary, so a long task can keep going.
Context engineering
Context engineering is the practice of choosing and maintaining everything that goes into a language model's context window during a task: instructions, retrieved documents, tool definitions, tool results and conversation history.
Context window
A context window is the maximum amount of text, counted in tokens, that a language model can consider at once, covering the system prompt, conversation, documents, tool results and the response it writes.
Retrieval-augmented generation
Retrieval-augmented generation (RAG) is a pattern where a system first searches a collection of documents for passages relevant to a question, then gives those passages to a language model so its answer draws on them.
Agent memory
Agent memory is the set of techniques that let an AI agent keep information beyond a single context window, such as facts about a user or past work, by storing it outside the model and loading it back when needed.

Sources