---
id: post/gerardorenteria-blog/14894
type: post
title: "💸The answer is the expensive part: controlling output tokens"
summary: Output tokens cost 6× more than input tokens in large language models, making it essential to control what Copilot writes rather than just what you send. The post explains hidden reasoning tokens and provides practical habits and prompt instructions to reduce output costs when using Copilot for AL development and code changes.
tier: community
language: en
tags:
  - copilot optimization
  - token costs
  - output control
  - al development
  - prompt engineering
  - reasoning tokens
  - cost efficiency
system: copilot
review:
  state: reviewed
  by: opus
  at: "2026-10-08T01:54:46.335Z"
  flags: []
generated:
  at: "2026-10-08T01:26:53.141Z"
  pipeline: 0.2.0
  prompts:
    extract-post: 1
  input_hash: 8c77d7ad3b226e812a519a871a941dc16471a3e7636752fcc555def88c15642e
evidence:
  - kind: blog
    url: https://gerardorenteria.blog/2026/06/21/%f0%9f%92%b8the-answer-is-the-expensive-part-controlling-output-tokens/
    title: "💸The answer is the expensive part: controlling output tokens"
    date: "2026-06-21"
    commit: null
    t: null
    quote: null
  - kind: blog
    url: https://gerardorenteria.blog/2026/06/21/%f0%9f%92%b8the-answer-is-the-expensive-part-controlling-output-tokens/
    title: "💸The answer is the expensive part: controlling output tokens"
    date: "2026-06-21"
    commit: null
    t: null
    quote: Each token the model writes costs us about 6× what each token we send does.
links:
  learn: []
  objects: []
  features: []
  topics: []
  localizations: []
  videos: []
  posts: []
  guidelines: []
post_id: "14894"
source_id: gerardorenteria-blog
source_name: Gerardo Renteria
url: https://gerardorenteria.blog/2026/06/21/%f0%9f%92%b8the-answer-is-the-expensive-part-controlling-output-tokens/
published_at: "2026-06-21T12:26:21.000Z"
author: Gerardo Renteria
full_text: false
words: 3613
quotes:
  - text: Each token the model writes costs us about 6× what each token we send does.
    why_it_matters: This fundamental price difference means controlling output is more impactful than controlling input, changing the entire optimization strategy
code_objects_mentioned: []
systems:
  - copilot
  - development
versions_mentioned: []
preview:
  embeddable: true
  frame_url: https://gerardorenteria.blog/2026/06/21/%f0%9f%92%b8the-answer-is-the-expensive-part-controlling-output-tokens/embed/
  image: https://gerardorenteria.blog/wp-content/uploads/2026/06/designer-1-1.png
  image_alt: null
  image_w: 1024
  image_h: 1024
  site_name: Gerardo Rentería Blog
  favicon: https://gerardorenteria.blog/wp-content/uploads/2020/11/cropped-wolficon.jpg?w=32
  probed_at: "2026-10-07T11:50:30.850Z"
---

# 💸The answer is the expensive part: controlling output tokens

[Read the post](https://gerardorenteria.blog/2026/06/21/%f0%9f%92%b8the-answer-is-the-expensive-part-controlling-output-tokens/) · Gerardo Renteria (Gerardo Renteria) · 2026-06-21 · 3613 words · tier community · reviewed (checked by Opus)

> Output tokens cost 6× more than input tokens in large language models, making it essential to control what Copilot writes rather than just what you send. The post explains hidden reasoning tokens and provides practical habits and prompt instructions to reduce output costs when using Copilot for AL development and code changes.

## Key points

- Output tokens are 6 times more expensive than input tokens because the model generates responses word-by-word, requiring multiple neural network passes
- Reasoning tokens (used for planning and self-correction in advanced models) are billed as output even when invisible on screen
- Surgical prompts requesting specific changes (e.g., modifying only a procedure instead of rewriting entire codeunits) can reduce costs by 19×
- Agent mode costs scale nonlinearly with iterations; use Ask mode for understanding, Plan for decisions, and Agent only for execution
- Permanent conciseness instructions and clear output format constraints (diff only, no prose, no alternatives) provide recurring high savings across all prompts

## Quotes

- "Each token the model writes costs us about 6× what each token we send does." (This fundamental price difference means controlling output is more impactful than controlling input, changing the entire optimization strategy)

## Context

- Features: Agent mode, Ask mode, Plan mode, Reasoning effort calibration, Output format constraints, Conciseness instructions, Prompt engineering for cost reduction

Source: Gerardo Renteria, community blog. Summary, key points and quotes are derived (CONTENT-NOTICE.md); read the original for the full text.