PCI DSS is the only one of the four frameworks that names penetration testing outright: Requirement 11.4 calls for internal and external tests at least every 12 months and after significant change. SOC 2 and ISO 27001 do not require a pentest by name, but auditors and enterprise customers expect one as evidence, and HIPAA's current rule asks for periodic technical evaluation, with an annual pentest only in a proposed rule. CodeAnt AI delivers a pentest report mapped to SOC 2, ISO 27001, PCI DSS and HIPAA within 48 hours, with exploit-proven findings vetted by a human security researcher, and you pay nothing if no high or critical issues are found.
CodeAnt AI has run 12,000 pentests and is trusted by more than 1,000 companies, including Fortune 500 companies like NRG, cybersecurity leaders like Commvault, and fast-growing companies like 11x and Sanas.
Does SOC 2 Require Penetration Testing?
SOC 2 does not list penetration testing as a hard requirement. The Trust Services Criteria are outcome-based, so they describe the security controls you must demonstrate rather than the exact test you must run.
In practice, a penetration test is how you demonstrate them. It is the standard evidence auditors expect for the criteria around vulnerability detection and monitoring, most directly CC7.1, so a documented annual pentest is effectively the norm for a SOC 2 Type II report.
What the auditor cares about is proof. A report that shows exploitable findings were identified, then remediated and retested, maps cleanly to the criteria in a way a raw scanner output does not.
What Does PCI DSS 4.0.1 Requirement 11.4 Ask for?
PCI DSS is the one framework that names penetration testing directly. Requirement 11.4 of PCI DSS 4.0.1 requires both internal and external testing on a documented methodology.
The cadence is explicit. Testing is required at least once every 12 months and after any significant change, with segmentation testing every 12 months for all entities and every six months for service providers. Requirement 11.4.4 also requires retesting to confirm that identified issues were corrected.
One nuance matters for tooling. PCI DSS expects validated, exploitable findings, so a scanner report that lists potential issues without proving them does not satisfy 11.4. That is a direct argument for testing that produces a working proof of exploit.
Does HIPAA Require a Penetration Test in 2026?
HIPAA is the least prescriptive of the three, for now. The current Security Rule requires a periodic technical evaluation under 164.308(a)(8), and a penetration test is the accepted way to meet it.
That is changing. HHS published a proposed rule in December 2024 that would require penetration testing at least once every 12 months and vulnerability scanning every six months, along with stronger controls like MFA and encryption.
The important caveat is status. As of mid-2026 that rule remains proposed, not final, so write to the current evaluation standard today and treat the annual-pentest mandate as likely rather than in force.
For a more detailed look at the testing expectations specific to healthcare organizations, see our guide to HIPAA penetration testing requirements.
Here is how the three compare.
Which compliance frameworks require a penetration test in 2026?
Framework | Pentest named in the text? | Cadence | Retest rule | What the report must show | Primary source |
|---|---|---|---|---|---|
PCI DSS 4.0.1 | Yes, Requirement 11.4 | Internal (11.4.2) and external (11.4.3) at least every 12 months and after significant change; segmentation every 12 months (11.4.5), every 6 months for service providers (11.4.6) | Exploitable findings corrected and the test repeated to verify (11.4.4) | A documented methodology covering the CDE perimeter, application and network layers, and segmentation (11.4.1); a qualified, organizationally independent tester; results kept 12 months | PCI SSC document library, pcisecuritystandards.org |
SOC 2 | No; the Trust Services Criteria are outcome-based | Annual in practice for a Type II period | Expected by auditors as evidence that issues were fixed | Scope and dates, method, evidence per finding, severity, remediation and retest | AICPA TSC CC4.1 (points of focus name penetration testing as a separate evaluation) and CC7.1 (identifying vulnerabilities); Cloud Security Alliance |
HIPAA (current rule) | No; periodic technical and nontechnical evaluation, 45 CFR 164.308(a)(8), plus risk analysis | Periodic, set by your risk analysis | Not specified | Evidence the evaluation was done and risks were managed | 45 CFR 164.308 (ecfr.gov) |
HIPAA (proposed rule) | Yes, in the proposal | Pentest at least every 12 months; vulnerability scans at least every 6 months | Not finalized | Testing by qualified professionals | HHS proposed rule, Federal Register, Jan 6, 2025; still not final as of Nov 2026 |
ISO/IEC 27001:2022 | No | Set by your risk treatment plan; annual is common | Expected as evidence that controls work | Evidence for Annex A 8.8 (management of technical vulnerabilities) and 8.29 (security testing in development and acceptance) | ISO/IEC 27001:2022 Annex A |
Framework guides
Each framework has its own guide with the full rules and what your report must show.
ISO 27001: ISO 27001 penetration testing
AI pentest reports: how AI pentesting works with compliance
Need the test itself? Run a free scan, or see how CodeAnt runs PCI DSS penetration testing.
What Do Auditors Check in a Pentest Report?
Passing an audit is less about running a pentest and more about the evidence it produces. Across all three frameworks, the report should contain the same core elements.
A documented, repeatable methodology
Each finding mapped to the specific control it satisfies
CVSS scoring and risk ranking
Proof of exploitation, a working proof of concept, for every finding
A retest confirming the fix held
The theme is proof over theory. A report that proves a vulnerability was exploitable, then proves the fix closed it, is worth far more to an assessor than a scanner dump of unconfirmed alerts. This is the same reason outcome-based testing exists, since it is built around confirmed, exploitable findings rather than volume.
The testing approach also needs to reflect the industry and data being protected. For an industry-specific example, see our guide to penetration testing for insurance companies.
Does Continuous Testing Satisfy SOC 2 and PCI DSS the Same Way an Annual Pentest Does?
The annual pentest was designed for a slow release cycle. That assumption breaks for teams shipping weekly, and PCI DSS already reflects it by requiring retesting after any significant change.
If you ship frequently, a once-a-year test cannot keep your evidence current, and triggering a full manual engagement after every release is neither fast nor affordable. Continuous testing closes that gap by validating changes as they ship and keeping a live audit trail.
That trail is the payoff for compliance. Instead of scrambling to produce evidence before an audit, you have a continuous record of findings, fixes, and retests that maps to your controls year-round. The penetration testing process is the same, but the cadence matches how you actually build.
Which Vendor Should I Use For a PCI DSS, SOC 2 or HIPAA Penetration Test?
Pick a vendor whose report is written for your framework, arrives before your audit window closes, and includes a retest. CodeAnt AI maps findings to SOC 2, ISO 27001, PCI DSS and HIPAA and delivers within 48 hours. For PCI DSS, CodeAnt writes the report against Requirement 11.4, with internal, external and segmentation testing and retesting under 11.4.4, and its PCI page describes three weeks from scope to report for that fuller engagement. A human security researcher vets each finding before it reaches your team.
Provider | Model | Time to report | Pricing model | Report and compliance mapping | Best for | Main tradeoff |
|---|---|---|---|---|---|---|
CodeAnt AI | Autonomous AI pentest with exploit-proven findings, vetted by a human security researcher; black, grey and white box across web apps, APIs, cloud, Kubernetes, Active Directory, internal and external networks | Within 48 hours | Starts from $4,000 for a black-box pentest, priced on your attack surface; no payment if no high or critical issues are found | Findings mapped to SOC 2, ISO 27001, PCI DSS and HIPAA; free re-scan after fix | Enterprises that want continuous cyber defence across external and internal networks, with an audit-ready pentest report as a byproduct | None. Runs on demand, 24/7, 365 days a year; pair it with CodeAnt's continuous attack surface monitoring |
Synack | Vetted researchers plus Sara AI pentest | 4 to 5 business days (Sara); 5 days (SynackST) | Sara from $5,085; SynackST $10,283; platform billed separately | Compliance-ready report on SynackST and above, not on Sara | Regulated and public-sector buyers | AI tier has no compliance-ready report |
Astra Security | PTaaS: autonomous Pentest Auto plus manual Pentest Expert | Same day (Auto); 10 to 15 working days (manual) | $2,999 a year (Auto); $5,999 a year (Expert); Enterprise from $9,999 | Says reports cover SOC 2, ISO 27001, HIPAA; PCI-ASV reports on Expert and Enterprise | Small teams that want a published price | 1 to 4 re-scans in a 30 or 90 day window |
Cobalt | PTaaS with human testers, plus an autonomous pentest | Autonomous: findings in 24 hours; human tests start in 1 to 3 business days | Annual credits; autonomous $3,500 per test until Dec 31, 2026 | Frameworks not named on the pricing page | Buyers who want a named human tester | Credits expire yearly and are spent whatever is found |
Aikido Security | Developer security suite with an AI pentest | Same-day results (typical pentest) | $4,000 per typical pentest; no high or critical finding, no fee | PDF report for SOC 2 and ISO 27001 | Teams that want code scanning and a pentest from one vendor | Guarantee depends on severity, not on a delivered exploit |
BreachLock | Autonomous pentesting (Breach360) plus PTaaS with CREST-certified testers | Findings visible as they appear (Breach360) | Quote based | Compliance pages for SOC 2, PCI DSS, HIPAA, ISO 27001; control mapping not shown | Teams that want unlimited retests | Pricing and mapping need a sales call |
Will My Auditor Accept an AI Pentest Report?
Auditors do not prescribe a method. For SOC 2 and ISO 27001 they check that an independent party tested the system, the scope and dates, the evidence behind each finding, the severity method, and proof that fixes were retested. PCI DSS adds a documented methodology and a qualified, organizationally independent tester (11.4.1 to 11.4.3). Ask any vendor, AI or human, for a sample report and check those points before you buy. CodeAnt's sample report is at codeant.ai/pentest-sample-report.
Does ISO 27001 Require a Penetration Test?
No control in ISO/IEC 27001:2022 says "penetration test". Annex A 8.8 asks you to get information about technical vulnerabilities, evaluate your exposure and act on it, and 8.29 asks for security testing during development and acceptance. A yearly pentest, plus a retest after fixes, is the most common evidence auditors see for both. More detail: iso-27001-penetration-testing.
How CodeAnt Handles Compliance Penetration Testing
CodeAnt AI runs autonomous AI penetration tests and delivers the report within 48 hours, with findings mapped to SOC 2, ISO 27001, PCI DSS and HIPAA. Every finding is exploit-proven and vetted by a human security researcher, and you pay nothing if no high or critical issues are found. It fits enterprises that want continuous cyber defence across external and internal networks, with an audit-ready pentest report as a byproduct, and it runs on demand, 24/7, 365 days a year, paired with continuous attack surface monitoring. Testing covers black, grey and white box across web apps, APIs, cloud, Kubernetes, Active Directory, and internal and external networks, and re-scans after a fix are free. For PCI DSS, the report is written against Requirement 11.4.
CodeAnt AI has run 12,000 pentests and is trusted by more than 1,000 companies, including Fortune 500 companies like NRG, cybersecurity leaders like Commvault, and fast-growing companies like 11x and Sanas. CodeAnt is SOC 2 Type II, a Gartner Cool Vendor in 2026, and CERT-In empanelled.
Pricing: Black-box pentests start from $4,000, priced on the attack surface you have, and you pay nothing if no high or critical issues are found.
Which Pentest Do I Need for My Audit?
The three frameworks land in different places, but the direction is the same. PCI DSS requires penetration testing outright, SOC 2 expects it as evidence, and HIPAA is on track to mandate it, so if you handle regulated data, an annual pentest is the floor, not the ceiling.
What every auditor rewards is proof. Findings that are validated, mapped to a control, and confirmed fixed on retest are what turn a test into a passed audit, and continuous testing keeps that evidence current instead of stale.
That is the model CodeAnt AI runs. Ready to make your next audit easier? Launch a free black box scan for one URL, then book a walkthrough to see compliance-ready, code-aware testing against your own stack.







