Prompt

A prompt is input supplied to a generative model or system to shape the task, context, constraints, or desired output.

Plan for this page

Understand one concept in plain language and connect it to nearby ideas.

You will leave with: It may contain a question, instruction, examples, source material, role information, or output format. The result also depends on system instructions, model, tools, retrieval, history, sampling, and product controls.
Time: 3–8 min
Before you begin: No prerequisite. Keep one real example in mind.
Do this now: Read the explanation, then compare the example with the common misunderstanding.

Prepared next stepPractice context and tool choices

PromptA prompt is input supplied to a generative model or system to shape the task, context, constraints, or desired output.

Context window
System instruction
Temperature

Plain explanation

It may contain a question, instruction, examples, source material, role information, or output format. The result also depends on system instructions, model, tools, retrieval, history, sampling, and product controls.

Example

“Using only the attached policy, list the three deadlines and quote the section supporting each” provides task, source boundary, format, and evidence request.

Why it matters

Clear context and constraints often improve usefulness, but prompt writing cannot guarantee truth, permission, privacy, or safe action. Verification and system design remain necessary.

Common misunderstanding

Misconception: a secret perfect phrase unlocks reliable intelligence. Correction: prompts influence behavior within the model and system’s actual capabilities and limits.

Related concepts

Follow relationships, not only the alphabet

Context windowA context window is the limited amount of tokenized information a model can consider together during one generation or inference.System instructionA system instruction is configuration context intended to guide a model’s behavior across a session or task, often with priority over ordinary user input.TemperatureTemperature is a generation setting that changes how strongly token sampling favors higher-probability options.

Practice context and tool choices

Sources and scope

What supports this object

Machine Learning GlossaryGoogle for Developers

Defines prompts in generative-model use; product-specific instruction hierarchies and behavior differ.

Open source

Loading page content...

…