The difference in plain words
A vulnerability assessment finds and ranks the known weaknesses across your systems, mostly with automated scanners, and does not try to exploit them. A penetration test has a tester attack the system with permission, exploit what it can, and report what an attacker could reach. The assessment tells you what might be wrong everywhere. The pentest proves what is wrong in the scope you chose.
The two are often sold together as VAPT (vulnerability assessment and penetration testing), and most compliance programs need both on different schedules. If an auditor, a customer, or PCI DSS has asked you for “a pentest”, a scan report will usually not answer the request. The sections below explain why, and how to check what a quote includes.
Vulnerability assessment vs penetration testing, side by side
| Vulnerability assessment | Penetration test | |
|---|---|---|
| What it is | Automated scanning of hosts, applications and cloud configuration, plus review and risk ranking of the results | An authorized attack that tries to exploit weaknesses and chain them into access |
| Question it answers | Which known weaknesses exist across our estate? | What can an attacker do from here? |
| Who does it | Your own team with a scanner, a managed service, or a PCI Approved Scanning Vendor for external card scans | An independent tester: a consultancy, a PTaaS provider, or a qualified internal team separate from the system owners |
| Depth | Wide and shallow: every asset, matched against known signatures and versions | Narrow and deep: a defined scope, including logged-in roles, business logic and multi-step attack paths |
| False positives | Common. Findings are inferred from version strings and responses, so someone has to confirm them | Few when each finding carries proof of exploitation. Ask whether findings are reproduced before they are reported |
| Typical cost | Low. Often bundled into tools you already pay for, or a per-asset subscription | Higher, because you pay for tester time. Most professional tests are budgeted at $10,000 to $30,000 |
| Frequency | Weekly to quarterly, and after significant changes | At least annually, and after significant changes |
| What auditors accept it for | Vulnerability management evidence, including PCI DSS requirement 11.3 scans | Security testing evidence: PCI DSS requirement 11.4, SOC 2 pentest controls, and customer security reviews |
The cost band for a pentest comes from published 2026 vendor and regulatory estimates collected in our penetration testing cost guide. Scope moves the price more than anything else.
Scan, assessment, pentest: the three terms
Vendors use these words loosely, so pin each one to a definition before you compare quotes. NIST's technical testing guide, SP 800-115, supplies two of them.
- Vulnerability scanning is “a technique used to identify hosts/host attributes and associated vulnerabilities” (NIST glossary, from SP 800-115). It is the tool run.
- Vulnerability assessment is the wider job: NIST defines it as a systematic examination of a system to determine whether its security measures are adequate and to identify deficiencies (NIST glossary). In practice it means scanning plus a person who removes false positives, adds business context, and ranks what is left.
- Penetration testing is security testing in which evaluators “mimic real-world attacks in an attempt to identify ways to circumvent the security features of an application, system, or network”. SP 800-115 adds that most penetration tests look for combinations of vulnerabilities that give more access than any single one (NIST glossary, from SP 800-115).
The dividing line is exploitation. A scan or an assessment reports that a weakness probably exists. A pentest tries to use it and records what happened.
What a vulnerability assessment finds and misses
A scanner sends requests to every host, port and page in its target list, reads the responses, and compares them with a database of known issues. That makes it good at a specific class of problem.
What it finds well
- Software running a version with a published CVE
- Missing operating system and library patches
- Exposed services, open management ports and default credentials
- Weak TLS settings, missing security headers and expired certificates
- Cloud misconfiguration such as public storage buckets
What it misses
A scanner has no idea what your application is supposed to do. It cannot tell that a customer in one tenant should not read another tenant's invoices, that a refund can be replayed, or that a support role can promote itself to admin. Those are access control and business logic flaws, and they make up much of a real pentest report. Our web vulnerability field guide has examples of each.
A scanner also reports each issue on its own. Three medium findings that chain into account takeover look, to a scanner, like three medium findings. And because many results are inferred from a version banner, a patched backport can show up as vulnerable. Someone has to triage every result before engineers spend time on it.
What a penetration test adds
A penetration test starts from a written scope and signed authorization, then attacks the targets the way an intruder would. The tester logs in as each role you provide, maps how the system works, and tries to cross the boundaries it is supposed to enforce. When something works, the report shows the request, the response and the impact, so an engineer can reproduce it.
- Proof. Each finding has evidence that the weakness can be exploited, which removes most of the triage work.
- Chains. The tester combines small issues into the attack path an intruder would take.
- Logic and authorization. Tenant isolation, role checks and workflow abuse need someone who understands the application.
- Evidence for others. The report, with scope, method, dates, rated findings and retest results, is what auditors and customers ask to see. Our penetration test report template lists every section a complete report carries.
The trade-off is coverage. A pentest spends its time on the scope you bought, so assets outside it get no attention until the next test. How much the tester knows at the start also changes what it finds: see black box vs gray box vs white box penetration testing.
One application, two reports
Take a fictional multi-tenant SaaS billing app with a public web front end and a REST API.
The vulnerability assessment reports
- An outdated JavaScript library with a known CVE
- A missing Content-Security-Policy header
- TLS 1.0 still enabled on a legacy subdomain
- An admin login page reachable from the internet
The penetration test reports
- A user in tenant A can download tenant B's invoices by changing the invoice ID in an API call
- The outdated library is present but its vulnerable function is never called, so it is rated low
- A password reset token stays valid after use, which combines with the exposed admin page into a takeover path
Both reports are useful. The assessment covers every host and keeps patching on schedule. Only the pentest finds the cross-tenant data leak, which is the finding a customer's security team would care about most.
What PCI DSS, SOC 2 and customers ask for
PCI DSS
PCI DSS v4.0.1 keeps the two activities in separate requirements. Requirement 11.3 covers vulnerability scanning: internal scans at least once every three months, and external scans at least once every three months by a PCI SSC Approved Scanning Vendor (ASV), plus rescans after significant changes. Requirement 11.4 covers penetration testing: internal and external tests at least once every 12 months and after significant changes, and segmentation testing where segmentation reduces scope. A passing ASV scan does not satisfy 11.4. The official text is in the PCI SSC document library, and our PCI DSS penetration testing guide covers which self-assessment questionnaires include the pentest rows.
SOC 2
SOC 2 does not prescribe either one. The AICPA Trust Services Criteria mention both only as points of focus, which are illustrative: the monitoring criterion (CC4.1) lists penetration testing as one kind of separate evaluation, and the system operations criterion (CC7.1) lists periodic vulnerability scans. Your auditor tests the controls you describe, so if they name quarterly scans and an annual pentest, you need dated evidence of both. Our SOC 2 penetration testing requirements guide covers scope and timing for Type I and Type II reports.
Enterprise customers
Common security questionnaires ask separately whether you scan for vulnerabilities and whether an independent party has tested the product. When a customer asks for a pentest report, they want to see scope, methodology, dates and findings with their remediation status. Sending a scanner export in its place tends to start another round of questions.
Which one do you need?
| Your situation | What to get |
|---|---|
| PCI DSS assessment coming up | Both. Quarterly internal and ASV scans under 11.3, and an annual internal and external pentest under 11.4 |
| SOC 2 audit, or an auditor asked for “a pentest” | A penetration test of the in-scope product, with retest results, plus the scanning your controls describe |
| Enterprise customer security review | A recent third-party pentest report or attestation letter for the product they will use |
| Large estate, no idea what is exposed | Start with a vulnerability assessment to build the asset list and clear the known issues, then pentest the systems that matter most |
| Major release, new auth flow or new API | A targeted pentest of the change before or right after launch |
| Between annual tests | Continuous scanning, with a retest of each fix as it ships |
For a SaaS company the usual answer is a scanner running all year and a pentest at least once a year. The pentest scope should match what you sell: web application penetration testing for the product, API penetration testing for public and partner APIs, network penetration testing for internet-facing and internal infrastructure, and cloud penetration testing for AWS, Google Cloud or Azure accounts.
How to spot a scan sold as a pentest
Some low-priced “pentests” are a scanner run with a branded cover page. Price alone does not tell you which one you are buying, so ask these questions before you sign.
Does the tester log in as each role you provide, and test what each role can reach?
Will the tester attempt exploitation, and does each finding include the request, response and impact?
Are findings reproduced before they are reported, or passed through from a tool?
Does the report state scope, methodology, test dates and the tester's independence?
Are retests included, and does the report record their results?
Will the provider sign an attestation letter your auditor or customer can rely on?
A “no” to the first three means you are looking at a vulnerability assessment. That can be the right purchase, but it should be priced and labelled as one.
How Helix Cyber runs a pentest
Helix Cyber sells penetration tests. AI agents test your web app, API, network or source code under human oversight, and a second agent reproduces every finding before it reaches your report. Anything that cannot be reproduced stays a candidate and is left out. Retests are free and unlimited, so each fix gets a recorded result.
A typical web app or API costs $3,000 for a black-box test, $4,000 for gray box, and from $7,000 for white box. You order, authorize and track the test through a platform. Our guide to penetration testing as a service (PTaaS) explains how that model compares with a consultancy engagement.
Frequently asked questions
What is the main difference between a vulnerability assessment and a penetration test?
A vulnerability assessment identifies and ranks known weaknesses, mostly with automated scanners, and stops there. A penetration test attempts to exploit weaknesses, including ones no scanner has a signature for, and reports what an attacker could reach. The assessment covers breadth; the pentest proves impact.
Is vulnerability scanning the same as a vulnerability assessment?
Not quite. Scanning is the automated step: a tool probes hosts and applications and matches what it sees against known issues. An assessment adds a person reviewing that output, removing false positives, and ranking what is left by risk to your business. Many vendors use the two terms interchangeably, so check which one a quote includes.
Does a vulnerability scan satisfy a penetration testing requirement?
Not under PCI DSS, which lists vulnerability scans (requirement 11.3) and penetration tests (requirement 11.4) as separate requirements. SOC 2 has no fixed rule, but auditors and enterprise customers who ask for a penetration test usually expect a report that shows attempted exploitation. A scanner export rarely answers that request.
How often should you run each one?
Run vulnerability scans often: PCI DSS asks for internal and external scans at least every three months and after significant changes, and many teams scan weekly or on every deploy. Penetration tests are usually annual and after significant changes, which is the PCI DSS minimum, with extra tests before major releases or customer reviews.
Do I need both a vulnerability assessment and a penetration test?
Most organizations with compliance obligations do. Scans catch new known vulnerabilities between tests at low cost. The pentest finds the access control, business logic and chained flaws that scanners cannot see, and produces the evidence auditors and customers ask for.
Which costs more?
A penetration test. Scanning is often included in tools you already pay for or sold as a per-asset subscription. A pentest pays for tester time, and most professional tests are budgeted in the $10,000 to $30,000 range. Helix Cyber prices a typical web app or API pentest at $3,000 black box, $4,000 gray box, or from $7,000 white box.
Sources
- NIST SP 800-115: Technical Guide to Information Security Testing and Assessment
- NIST CSRC glossary: vulnerability scanning
- NIST CSRC glossary: vulnerability assessment
- NIST CSRC glossary: penetration testing
- PCI Security Standards Council document library (PCI DSS v4.0.1)
- AICPA 2017 Trust Services Criteria (with revised points of focus, 2022)
The SaaS example is fictional. The decision table is Helix's practical reading of these sources; confirm requirements for your own assessment with your Qualified Security Assessor (QSA) or auditor.