---
schema: 1
type: term
title: mega prompt
description: >-
  A **mega prompt** is a single, large prompt that bundles many instructions,
  requirements, and sub-tasks into one comprehensive request to an AI model,
  rather than splitting the work across multiple smaller prompts.
canonical: 'https://liquidgold.cc/channel/everyinc/explore/term/mega-prompt-9e0c2d15e033'
markdown: >-
  https://liquidgold.cc/channel/everyinc/explore/term/mega-prompt-9e0c2d15e033.md
llms: null
parent: 'https://liquidgold.cc/channel/everyinc/llms.txt'
channel: everyinc
updated: '2026-10-23T22:12:35.346Z'
language: null
children: []
---


# mega prompt

Canonical: https://liquidgold.cc/channel/everyinc/explore/term/mega-prompt-9e0c2d15e033

Parent: [Channel index](https://liquidgold.cc/channel/everyinc/llms.txt)

## Definition

A **mega prompt** is a single, large prompt that bundles many instructions, requirements, and sub-tasks into one comprehensive request to an AI model, rather than splitting the work across multiple smaller prompts.

In this channel's material, a **mega prompt** is a single, very large prompt — one that packs many instructions, requirements, and sub-tasks into a comprehensive request handed to an AI model at once, rather than working through a series of smaller back-and-forth prompts.Mega prompts come up in the context of the channel's live model tests and vibe checks [^8a401be3-f56a-5934-a449-63d1e01a151d], where the Every team evaluates how flagship models handle substantial, structured workloads. When a model is put through a mega prompt, you are effectively exploring how it copes with complexity in one shot: following many constraints at the same time, keeping earlier requirements in mind while fulfilling later ones, and organizing a long response around all of them at once. That makes it a useful probe when you want to understand a model's real working behavior — not just whether it answers a simple question well.In side-by-side comparisons, a well-constructed mega prompt exposes differences between models that short prompts hide: instruction-following discipline, length control, and how gracefully a model degrades when the request gets demanding. Keep it in your toolkit as one of the structured tests you can reuse when you run your own model comparisons.- A **mega prompt** is one large, comprehensive prompt combining many instructions and sub-tasks into a single request. - It is used in the channel's model vibe checks as a way to explore how models handle complex, multi-part workloads. - It is a reusable test for spotting real differences between models in side-by-side comparisons.