---
title: "Build Your Own Benchmark: A 30-Question Test for the Work You Actually Do"
dek: "Public leaderboards measure someone else's job. This step-by-step guide helps you build a small, honest test that tells you which model is good at yours."
author: "Lena Hartmann"
author_title: "Research Editor"
category: "Models & Research"
format: guide
tags: ["Research", "Work", "Evaluation", "Workflows"]
published: 2026-09-30T23:25:07.284Z
updated: 2026-09-30T23:25:07.284Z
reading_minutes: 4
url: "http://localhost:3000/article/build-your-own-benchmark-30-questions"
ai_disclosure: assisted
ai_disclosure_label: "AI-assisted"
ai_notes: "Lena used a model to brainstorm edge-case examples for the task list; the rubric, the Fernhill case study and all recommendations are her own."
members_only: true
summary:
  - "Collect 30 real tasks: 10 routine, 10 tricky, 5 edge cases and 5 where the right move is to ask or refuse."
  - "Score each answer from 0 to 2 against a written checklist, for a maximum of 60."
  - "Run the tests blind so brand loyalty doesn't skew scores, then read the failures as well as the totals."
  - "Re-run the set every quarter and retire questions that every model now aces."
publisher: "Ideabro"
---

# Build Your Own Benchmark: A 30-Question Test for the Work You Actually Do

*Public leaderboards measure someone else's job. This step-by-step guide helps you build a small, honest test that tells you which model is good at yours.*

Public leaderboards are useful the way restaurant reviews are useful. They tell you what's popular and roughly what's good, but they can't tell you whether you'll like the soup.

Your work is specific. A model that tops a reasoning leaderboard might still mangle your company's tone, misread your spreadsheets or refuse a perfectly ordinary request in your field. The only way to know is to test it on your own tasks, and a useful personal benchmark takes an afternoon, not a research lab.

Here's how to build one with thirty questions.

> This is a members-only story. The full text is available to signed-in readers at http://localhost:3000/article/build-your-own-benchmark-30-questions

---

Source: [Build Your Own Benchmark: A 30-Question Test for the Work You Actually Do](http://localhost:3000/article/build-your-own-benchmark-30-questions) — Ideabro. AI-assisted: Written by a person with AI help for research, outlining or editing.