Skip to content

Instantly share code, notes, and snippets.

@DenverCoder1
Last active June 2, 2026 08:50
Show Gist options
  • Select an option

  • Save DenverCoder1/7fc11360534ed93d609e7d64ba0d7515 to your computer and use it in GitHub Desktop.

Select an option

Save DenverCoder1/7fc11360534ed93d609e7d64ba0d7515 to your computer and use it in GitHub Desktop.
GitHub Copilot Token-Saving Checklist

GitHub Copilot Token-Saving Checklist

Use this checklist to keep GitHub Copilot usage efficient and reduce the chance of hitting limits. GitHub now distinguishes between session and weekly usage limits, and both depend on token consumption and the model’s multiplier.

https://github.blog/news-insights/company-news/changes-to-github-copilot-individual-plans

Core rules

  • Start a new chat when the task changes.
  • Keep each prompt focused on one task.
  • Give complete requirements up front to avoid revisions.
  • Use inline suggestions and next-edit suggestions before chat.
  • Share only the relevant code, not entire files.
  • Use a lighter model for simple work.
  • Use Auto when you want Copilot to pick an efficient model.
  • Avoid parallel or repeated requests unless necessary.
  • Ask for “code only” when you do not need explanations.
  • Clean up noisy context before prompting.

What to do first

  • Try inline completion for small edits.
  • Use the simplest model that can solve the task.
  • Separate unrelated work into separate chats.
  • Bundle related questions into one prompt instead of several follow-ups.

Prompting habits

  • State the goal clearly.
  • Include constraints, edge cases, and output format.
  • Paste only the function, snippet, or error that matters.
  • Avoid vague prompts like “fix this” or “make it better.”
  • Prefer one strong prompt over multiple weak ones.

When limits are near

  • Switch to a smaller-multiplier model for simpler tasks.
  • Reduce parallel workflows and broad multi-file requests.
  • Wait and retry if you hit a temporary limit.
  • Check usage in your IDE or Copilot UI if available.
  • Upgrade only if your normal workflow truly needs more capacity.

Rule of thumb

  • Use chat for thinking.
  • Use inline suggestions for drafting.
  • Use new chats for new problems.
  • Use Auto or a lighter model when quality demands are modest.

One-line version

  • Fresh chat, focused prompt, small context, inline first, light model.

GitHub’s documentation says usage limits are temporary guardrails, and GitHub recommends checking usage, adjusting request patterns, changing models, or waiting when limits are hit.

https://docs.github.com/en/copilot/concepts/usage-limits

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment