ModelRefs / Stress Testing (LLMs) — AI Glossary
Stress Testing (LLMs) — AI Glossary
Evaluating model behavior at extreme inputs—maximum context length, adversarial prompts, edge-case formats—to identify failure boundaries.
Overview
LLM stress testing: fill context to maximum length and verify quality doesn't degrade (RULER), adversarial prompts (prompt injection, jailbreaks), malformed JSON tool schemas, very long tool results, empty inputs, and inputs in unexpected languages. Essential QA before production deployment, especially for agent workflows with external input.
Reference details
| Topic | ecosystem |
|---|---|
| Last reviewed | 2026-06-24 |
Related terms
Continue your research
Use these connected ModelRefs sections to compare alternatives, inspect implementation paths, and review the evidence and governance boundaries relevant to Stress Testing (LLMs) — AI Glossary.
Frequently asked questions
What is Stress Testing (LLMs)?
Evaluating model behavior at extreme inputs—maximum context length, adversarial prompts, edge-case formats—to identify failure boundaries.
What concepts are related to Stress Testing (LLMs)?
Closely related concepts include load testing, adversarial attack, evaluation harness.