Assessment rubric builder with clear performance levels

You are an assessment specialist. Build a rubric for the task below.

Task students complete: [e.g. "a 1,000-word persuasive essay on a local issue"]
Learners: [AGE OR LEVEL]
What the task is meant to assess: [SKILLS OR OBJECTIVES]
Number of performance levels: [e.g. "4: beginning, developing, secure, excellent"]
Total points or grading scale, if any: [e.g. "out of 20"]
Things I do NOT want to grade here: [e.g. "handwriting", "spelling, assessed elsewhere"]

Produce:
1. Four to six criteria. Each must assess one thing only, with no overlap between criteria.
2. A table: rows are criteria, columns are levels. Each cell describes what the work shows, using observable features, not vague words like "good" or "poor".
3. Points or weightings for each criterion, with a one-line reason for the weighting.
4. A student-friendly version of the same rubric, written as "I can..." statements.
5. Two common borderline cases and how to decide between the levels.

Rules: each level must differ from the next in a way a marker can see in the work. If my objectives are unclear, ask me before writing the rubric.

Example (excerpt):
| Criterion | Developing | Secure | Excellent |
| Use of evidence | Gives one example, not linked to the claim | Gives two relevant examples and explains how they support the claim | Uses varied evidence, explains it and addresses a counter-example |

Weighting: Use of evidence 30%, because it is the main skill this unit teaches.
Borderline case: strong evidence but no counter-argument. Award Secure, not Excellent, since addressing objections is the distinguishing feature.

When to use it

Use this prompt before you set a task, not after the work comes in. A rubric shared in advance tells students what quality looks like, and it makes marking quicker and more consistent across a class or between colleagues. It works for school essays, university projects, presentations, lab reports and workplace training assessments.

How to customize it

• [WHAT THE TASK IS MEANT TO ASSESS] should come from your objectives or curriculum. The rubric can only be as clear as the goals behind it.
• [NUMBER OF LEVELS] changes how fine the distinctions are. Four levels avoid a “safe middle” option; three are easier to apply quickly.
• [GRADING SCALE] lets the model propose weightings that add up correctly. Check the arithmetic yourself.
• [THINGS I DO NOT WANT TO GRADE] prevents criteria that punish students for skills outside the task, a frequent source of unfair marks.

Why it works

Weak rubrics fail in two ways: criteria overlap, so one weakness is penalized twice, and descriptors use relative words such as “good” or “excellent” that markers interpret differently. The prompt forbids both. It requires observable features in each cell and a visible difference between levels.

The borderline cases add a practical step that most rubric templates skip. Deciding in advance how to treat typical in-between work is what makes grading consistent. The same principle is used to judge model outputs, as in our LLM-as-judge rubric prompt.

Tips and variations

• Test the rubric on two or three real pieces of past work before using it. If markers disagree, tighten the descriptors.
• Ask for a single-point rubric (one column describing “meets the standard”, with space for comments) for formative feedback.
• Share the student-friendly version at the start of the task and ask students to self-assess before submitting.
• Keep final marking decisions with you; use the rubric to support judgment, not replace it.
• Pair it with our lesson plan prompt so objectives and assessment match.
• Ask the model to grade one anonymized sample against the rubric and explain each level it chose. If its reasoning surprises you, the descriptors probably need sharper wording before you use them with a class.