How AWS Scores Your Exam: The 100–1,000 Scale Explained

Every AWS certification exam reports your result as a scaled score between 100 and 1,000 — not a raw percentage of questions answered correctly, and not the same pass mark across every exam. Both of those surprise people who assume "score" means "percent correct." Here's what AWS's own exam guides actually say about how it works.

It's a scaled score, not a raw percentage

AWS states this identically across its exam guides: "Your results for the exam are reported as a scaled score of 100–1,000... Scaled scoring models help equate scores across multiple exam forms that might have slightly different difficulty levels." That last part is the actual reason the scale exists. AWS runs multiple different versions of the same exam concurrently — different candidates sitting "SAA-C03," for instance, may see a different specific set of questions than each other, drawn from a large question pool. If two exam forms differ slightly in difficulty, a raw percentage would unfairly reward whoever happened to get the easier form. Scaling adjusts for that, so a passing score means the same thing regardless of which specific set of questions you personally saw.

The pass mark isn't the same number across every exam

This is a common point of confusion, especially for anyone assuming certifications are internally consistent: the pass mark varies by exam level.

Exam level	Example	Passing score
Foundational	Cloud Practitioner, AI Practitioner	700
Associate	Solutions Architect Associate, Developer Associate, Data Engineer Associate, ML Engineer Associate	720
Professional	Solutions Architect Professional, DevOps Engineer Professional	750
Specialty	Advanced Networking Specialty	700

A 700 on a Foundational exam and a 700 on a Specialty exam both represent "pass," but they're not directly comparable achievements — each exam's scale is calibrated to its own difficulty and content, not to a universal 1,000-point ladder shared across every certification.

Compensatory scoring: you don't need to pass every section

Every AWS exam guide describes the same scoring model: "The exam uses a compensatory scoring model, which means that you do not need to achieve a passing score in each section. You need to pass only the overall exam." Practically, this means a genuinely strong performance in your best domain can offset a weaker one elsewhere in the same exam — you're not required to individually clear the bar in Security, then separately clear it in Resilience, then separately in every other domain. It's one combined score across the whole exam.

Your score report does include "a table of classifications of your performance at each section level" — but AWS explicitly cautions candidates to "use caution when you interpret section-level feedback," describing it as general guidance rather than a precise diagnostic of exactly which questions you missed.

Unscored questions are mixed in, and you can't tell which ones they are

Every AWS exam includes a batch of unscored questions alongside the scored ones — 15 unscored on most Foundational and Associate exams, 10 on Professional-level exams — which AWS uses to trial new questions for future exam versions. These are deliberately not identified anywhere on the exam. There's no way to spot them and skip extra effort on them; the only sound strategy is treating every question as if it counts, since roughly 15-20% of what you're answering genuinely doesn't affect your result, and you have no way of knowing which ones.

What this means practically

Don't over-interpret a single missed question or a weak-looking section on a score report as the reason you failed — the compensatory model means your overall result is what matters, not any individual section in isolation. And don't compare a scaled score directly across different exam levels as if they sit on one shared ladder — a 750 on a Professional-level exam (comfortably above its 750 pass mark) and a 750 on an Associate-level exam (comfortably above its lower 720 pass mark) reflect entirely separate calibrations, not the same underlying difficulty or the same margin of achievement.

If you're preparing and want practice that reflects this properly, our certification question banks are structured around each exam's real domain weightings, so your practice sessions build toward the actual scoring model rather than a flat right-or-wrong count.

JavaScript is required to use this application.

Please enable JavaScript in your browser to access the full NestedCerts platform.

Contact: support@nestedcerts.com