How the AI works

Context Rot: Why Longer AI Conversations Get Less Reliable

> ANSWER[AEO]

Feeding an AI model more input doesn't just risk hitting a hard limit — it degrades output quality well before that limit, even on simple tasks. Researchers at Chroma tested 18 leading models in 2025 and found every single one got measurably less reliable as input length increased, a pattern they named "context rot."

What did the study actually test?

Chroma Research — Kelly Hong, Anton Troynikov, and Jeff Huber — ran controlled experiments across GPT-4.1, the Claude 4 family, Gemini 2.5, Qwen3, and other current frontier models, varying input length and measuring accuracy on retrieval and reasoning tasks. The headline finding: reliability decreased significantly with longer inputs, and the decline wasn't smooth or uniform — it depended on how similar the target information was to surrounding "distractor" text, how the input was structured, and what kind of task was being asked.

Notably, this shows up even on tasks that sound trivial, like retrieving a fact or repeating back a piece of text — the kind of task where you'd expect a model to be reliable regardless of how much surrounding context there is.

How much does accuracy actually drop, and at what length?

The degradation isn't a cliff at the advertised limit — it starts much earlier. One analysis summarising the Chroma findings and related work describes a 200,000-token window showing serious accuracy loss by around 50,000 tokens of input, well under a quarter of its rated capacity.

Separate work applying similar methodology found accuracy drops of 30–50% appearing well before models reached their documented context limits, with the steepest decline typically in the higher end of the tested range.

Why does coherent, well-organised text sometimes perform worse?

This is the part of the research that surprises people. Across all 18 models tested, performance was better on shuffled, less coherent input than on logically structured documents of the same length. The likely explanation: a well-organised, plausible-sounding passage is a more convincing distractor to a model than a jumbled one — it reads as more relevant even when it isn't, which pulls attention away from the actual answer.

That finding matters practically: it means "just clean up your prompt formatting" isn't a reliable fix for context rot on its own.

What does this mean for how AI products should be built?

A few conclusions follow reasonably from the current evidence, without overstating it.

Our companion piece covers why AI forgets the middle of long conversations in detail, and how AI memory actually works covers the retrieval architecture that exists to avoid both problems.

Key takeaways

  • > Context rot is the drop in output quality as input grows, well before the advertised context limit.
  • > All 18 frontier models Chroma tested in 2025 showed the degradation.
  • > It appears even on trivial tasks like retrieving a fact or repeating text back.
  • > Coherent, well-structured input performed worse than shuffled input of the same length.
  • > Tidier formatting is not a fix; retrieving less is.

Sources

Frequently asked questions

What is "context rot"?
Context rot is the measurable drop in an AI model's output quality as the amount of input text grows, even when the model is nowhere near its maximum context limit.
  • Where the name comes from: a 2025 study from Chroma Research tested 18 frontier models, including GPT-4.1, Claude 4, and Gemini 2.5.
  • Every one of them showed this degradation as input length increased.
Is context rot the same thing as "lost in the middle"?
They're related but not identical: lost in the middle describes where in the input a model loses accuracy, while context rot is the broader finding that overall reliability declines as input grows.
  • Lost in the middle is positional — the middle of the input, specifically. Why AI forgets the middle of long conversations covers that effect in detail.
  • Context rot is driven by several compounding factors, of which positional bias is just one.
Does organizing your input more clearly help?
Not necessarily, and this is the counterintuitive part: the Chroma study found models performed better on shuffled, less-structured input than on coherent, logically-organized documents of the same length.
  • The likely reason: well-organized text can read as more convincing to the model, making it a more effective distractor rather than an easier reference point.
  • What does help is sending less — see how AI memory actually works for the retrieve-instead-of-resend alternative.