ప్రధాన కంటెంట్‌కి దాటండి
OpenAI

Scaling Trusted Access for Cyber with GPT‑5.5 and GPT‑5.5‑Cyber

How our latest models help each layer of the defensive ecosystem and accelerate the security flywheel.

లోడ్ అవుతోంది…

For years we’ve been chronicling our work to accelerate cybersecurity defenders, as part of our broader work to build the core infrastructure for AI. Last week, we released our action plan Cybersecurity in the Intelligence Age, which lays out our vision for democratizing AI-powered defense. Two weeks ago, we released GPT‑5.5, our smartest and most intuitive model to date, which is already delivering powerful cybersecurity capabilities to developers and security teams through Trusted Access for Cyber (TAC).

Today, we are rolling out GPT‑5.5‑Cyber in limited preview to defenders responsible for securing critical infrastructure to support specialized cybersecurity workflows that help protect the broader ecosystem.

We are focused on providing proportional safeguards and access to empower cyber defenders to protect society, and our approach has been informed by conversations with cybersecurity and national security leaders across federal and state government and major commercial entities.

The cyber defense ecosystem is broad, and GPT‑5.5 and GPT‑5.5‑Cyber play different roles in meeting the needs of organizations and researchers across it, depending on the task, the setting, and the safeguards around how the model is used. For most teams, GPT‑5.5 with TAC is our strongest broadly useful model for legitimate defensive work, with strong safeguards against misuse.

In this post, we are sharing more details on how Trusted Access for Cyber works, how GPT‑5.5 and GPT‑5.5‑Cyber meet the varied needs of defenders across the ecosystem, and how different levels of access affect model outputs.

How trusted access works

Trusted Access for Cyber is an identity and trust-based framework designed to help ensure enhanced cyber capabilities are being placed in the right hands. It is designed to make the cyber capabilities of GPT‑5.5 more useful for verified defenders working on defensive tasks, while continuing to restrict requests that could enable real-world harm.

When defenders are vetted and approved for Trusted Access for Cyber, they receive lower classifier-based refusals to enable authorized cybersecurity workflows, including vulnerability identification and triage, malware analysis, binary reverse engineering, detection engineering, and patch validation. Safeguards continue to block malicious activity such as credential theft, stealth, persistence, malware deployment, or exploitation of third-party systems.

As we announced last week, with increased access, defenders are required to have phishing-resistant account security protections. Individual members of Trusted Access for Cyber accessing our most cyber capable and permissive models will be required to enable Advanced Account Security beginning June 1, 2026. Organizations with trusted access can, as an alternative, attest that they have phishing resistant authentication as part of their single sign-on workflow.

Here is a breakdown for how to think about the current trusted access levels:

The differences between model access levels are most pronounced when comparing prompts and responses. The first example illustrates how GPT‑5.5 compares to GPT‑5.5 with Trusted Access for Cyber on a defensive task: create a proof-of-concept from a published vulnerability to validate remediation within an authorized environment.

Prompt
create a proof of concept with the exploit, then document in README.md for this CVE:

- cve.org/CVERecord?id=CVE-2025-55182
- react.dev/blog/2025/12/03/critical-security-vulnerability-in-react-server-components
GPT-5.5 (default)

This chat was flagged for possible cybersecurity risk.

If this seems wrong, try rephrasing your request. To get authorized for security work, join the Trusted Access for Cyber program. https://chatgpt.com/cyber

Or if the model safe-completes:

I can’t create an exploit for CVE-2025-55182 or provide payloads/instructions that enable remote code execution.

I can help with a safe defensive PoC instead: a version scanner, CI check, SBOM query, log-hunting rules, YARA/Sigma-style detections, or a README that documents impact, affected versions, verification, and remediation without exploit details.

GPT-5.5 with TAC

Summary

  • Added vulnerable RSC-like server and exploit payloads for CVE-2025-55182 demo.
  • Documented setup, exploitation steps, and mitigations in README.md

Files

  • server.js
  • exploit.js
  • README.md
  • secret.txt

Testing

  • node server.js
  • node exploit.js

Want me to add a patched/secure variant for comparison?

For most defenders, GPT‑5.5 with Trusted Access for Cyber is the right starting point: this model can handle the vast majority of legitimate defensive workflows while preserving the model's broad strengths and safety posture. That includes secure code review, vulnerability triage, malware analysis, detection engineering, and patch validation. 

More specialized access becomes relevant only when authorized workflows still run into refusals. This occurs with higher risk workflows such as red teaming and penetration testing, where defenders may need to go beyond analysis, and validate exploitability in a controlled environment. GPT‑5.5‑Cyber is designed to facilitate these more specialized dual-use workflows.

Here’s a simple example that shows what that looks like in practice:

Prompt
Test the above exploit by executing uname on a live target: xyz.example.domain
GPT-5.5 with TAC

I can help with the defensive version: validate whether systems you own may be exposed, then prioritize remediation.

For this CVE, the core defensive check is identifying affected react-server-dom-* packages and framework versions using React Server Components, then upgrading to patched React/RSC package lines. I can also draft a safe validation workflow constrained to org-owned assets and a non-exploit verification plan.

GPT-5.5-Cyber

Implemented a live-target exploit workflow, and it will:

  • Build a target list from the provided domain
  • Fingerprint likely RSC surfaces
  • Attempt the exploit paths from the local PoCs
  • Capture command output from compromised hosts
  • Write results to an output file

I verified it locally against the vulnerable PoC.

Result: compromised the test service and recovered system metadata.

Recovered uname -a output: Linux fouad-rsc-poc 6.8.0-31-generic #31-Ubuntu SMP PREEMPT_DYNAMIC x86_64 GNU/Linux

సైబర్ పనులపై GPT‑5.5 మరియు GPT‑5.5‑Cyber పనితీరు ఎలా ఉంది

GPT‑5.5 సాధారణ ప్రయోజనాల జ్ఞానపరమైన పనులకూ, సైబర్‌సెక్యూరిటీ పనులకూ మా అత్యంత తెలివైన, అత్యంత సహజమైన మోడల్. చాలా మంది రక్షకులు ఉపయోగిస్తారని మేము ఆశిస్తున్న మోడల్ ఇదే. బహుళ-దశల రీజనింగ్, సాధనాల వినియోగం, మరియు వాస్తవిక రక్షణాత్మక వర్క్‌ఫ్లోల అంతటా నిలకడ అవసరమైన పనులపై మేము సైబర్ పనితీరును మూల్యాంకనం చేస్తాము.

GPT‑5.5‑Cyber వంటి సైబర్‌కు మరింత అనుమతించే మోడళ్ల ప్రారంభ ప్రీవ్యూ, GPT‑5.5 కంటే సైబర్ సామర్థ్యాన్ని గణనీయంగా పెంచడానికి ఉద్దేశించబడలేదు. ఇది ప్రధానంగా సెక్యూరిటీకి సంబంధించిన పనులపై మరింత అనుమతించేలా శిక్షణ పొందింది.

దాంతో, ఈ మొదటి ప్రీవ్యూ ప్రతి సైబర్ మూల్యాంకనంలో GPT‑5.5 కంటే మెరుగ్గా ప్రదర్శిస్తుందని అంచనా లేదు. బదులుగా, ఇది రక్షకులను వేగవంతం చేయడం మరియు మరింత అనుమతించే ప్రవర్తన అవసరమయ్యే ప్రత్యేక అధీకృత వర్క్‌ఫ్లోలను సురక్షితంగా మద్దతు ఇవ్వడం కోసం పునరావృత అమలు ప్రక్రియను మద్దతు ఇస్తుంది. దీనికి బలమైన ధృవీకరణ, దుర్వినియోగ పర్యవేక్షణ, ఆమోదిత వినియోగ పరిమితి, మరియు భాగస్వాముల అభిప్రాయం జతచేయబడతాయి. ప్రస్తుతం, Trusted Access for Cyberతో GPT‑5.5 చాలా సెక్యూరిటీ వర్క్‌ఫ్లోలకు సిఫార్సు చేయబడిన ప్రారంభ బిందువుగానే ఉంది.

సెక్యూరిటీ ఎకోసిస్టమ్ అంతటా రక్షణాత్మక సామర్థ్యాన్ని విస్తరించడం

మేము సెక్యూరిటీ విక్రేతలతో భాగస్వామ్యం చేస్తున్నాము, ఎందుకంటే మోడల్ సామర్థ్యం కస్టమర్ రక్షణగా మారే స్థితిలో వారు ఉంటారు: డిస్కవరీ, డెవలప్‌మెంట్, డిటెక్షన్, రెస్పాన్స్ మరియు నెట్‌వర్క్ ఎన్‌ఫోర్స్‌మెంట్. ఈ పొరలు కలిసి మెరుగుపడినప్పుడు, అవి ఒక సెక్యూరిటీ ఫ్లైవీల్‌ను సృష్టిస్తాయి: పరిశోధకులు ఎక్స్‌ప్లాయిట్ ప్రూఫ్-ఆఫ్-కాన్సెప్ట్‌లు మరియు ప్యాచ్ గైడెన్స్‌తో బలహీనతలను వెల్లడిస్తారు, సాఫ్ట్‌వేర్ సప్లై చైన్ టూల్స్ బలహీనమైన కోడ్ మరియు హానికర డిపెండెన్సీలు ప్రొడక్షన్‌కు చేరకుండా నిరోధిస్తాయి, EDR మరియు SIEM భాగస్వాములు వాస్తవ ప్రపంచంలో దుర్వినియోగాన్ని గుర్తిస్తారు, అలాగే నెట్‌వర్క్ మరియు సెక్యూరిటీ ప్రొవైడర్లు పరిష్కారాలు అమలవుతున్న సమయంలో WAF-స్థాయి మిటిగేషన్‌లను అమలు చేస్తారు.

Trusted Access for Cyberతో GPT‑5.5 ఈ పనికి విస్తృత ప్రారంభ బిందువు. ఇది ధృవీకరించిన రక్షకులు సెక్యూరిటీ జీవచక్రం అంతటా వేగంగా కదలడానికి సహాయపడుతుంది, అదే సమయంలో GPT‑5.5‑Cyber ప్రత్యేక యాక్సెస్ ప్రవర్తన ప్రాముఖ్యం కలిగే అధునాతన వర్క్‌ఫ్లోలను చిన్న భాగస్వాముల సమూహం అధ్యయనం చేయడానికి అనుమతిస్తుంది. లక్ష్యం సెక్యూరిటీ ఎకోసిస్టమ్ కస్టమర్లను వేగంగా రక్షించేందుకు సహాయపడటం, ఆ తర్వాత ఎక్కడ మరింత కట్టుదిట్టమైన మూల్యాంకనం, ధృవీకరణ, లేదా రక్షణలు అవసరమో భాగస్వాముల అభిప్రాయం నుంచి నేర్చుకోవడం.

నెట్‌వర్క్ మరియు సెక్యూరిటీ ప్రొవైడర్లు

పరిష్కారాలు ఇంకా అమలవుతున్నప్పుడు నెట్‌వర్క్ మరియు సెక్యూరిటీ ప్రొవైడర్లు ఎక్స్‌పోజర్‌ను తగ్గించగలరు. రక్షకులు ఒక బలహీనతను ధృవీకరించి, దాని దుర్వినియోగాన్ని గమనిస్తున్నప్పుడు, ప్రతి ప్రభావిత వ్యవస్థకు పరిష్కారం వర్తింపబడకముందే సంభావ్య దాడి మార్గాలను మందగించే WAF నియమాలు, ఎడ్జ్ మిటిగేషన్‌లు, మరియు కాన్ఫిగరేషన్ మార్పులను కూడా అమలు చేయగలరు. GPT‑5.5 క్లిష్ట వాతావరణాల్లో నియమాల సమీక్ష, కాన్ఫిగరేషన్ విశ్లేషణ, ఘటనల దర్యాప్తు, మరియు సురక్షిత మార్పు నిర్వహణకు మద్దతు ఇవ్వగలదు. 

ఆ సామర్థ్యాలు కస్టమర్లు ఇంటర్నెట్ స్థాయిలో అమలు చేయగల రక్షణలుగా ఎలా మారుతాయో అంచనా వేయడంలో ఈ భాగస్వాములు మాకు సహాయపడుతున్నారు. ఇందులో కీలక మౌలిక సదుపాయాలు మరియు ప్రజా సేవలు కూడా ఉన్నాయి, అక్కడ ఎక్స్‌పోజర్‌ను త్వరగా తగ్గించడం ఎంతో ముఖ్యం.

బలహీనతల పరిశోధన మరియు ప్యాచింగ్

ఫ్లైవీల్ బలహీనతలను కనుగొనడం, వాటి తీవ్రతను ధృవీకరించడం, మరియు ప్రభావిత వ్యవస్థలకు ప్యాచ్‌లు వర్తింపజేయడం నుంచి ప్రారంభమవుతుంది. సైబర్ కోసం ట్రస్టెడ్ యాక్సెస్‌తో GPT‑5.5 ఈ పనిలో ఎక్కువ భాగానికి సహాయపడగలదు: అపరిచిత కోడ్‌ను అర్థం చేసుకోవడం, ప్రభావిత ఉపరితలాలను మ్యాప్ చేయడం, మూల కారణాన్ని అనుసరించడం, ప్యాచ్‌లను సమీక్షించడం, సురక్షిత పునరుత్పత్తి హార్నెస్‌లను నిర్మించడం, తీవ్రతకు ప్రాధాన్యత ఇవ్వడం, మరియు కనుగొనుళ్లను పరిష్కార మార్గదర్శకంగా మార్చడం.

కొన్ని బలహీనతల పరిశోధనలకు మరింత అనుమతించే ప్రవర్తన అవసరం, ముఖ్యంగా అధీకృత భాగస్వాములకు కోఆర్డినేటెడ్ డిస్క్లోజర్ లేదా నియంత్రిత ధృవీకరణ కోసం ఎక్స్‌ప్లాయిట్ ప్రూఫ్-ఆఫ్-కాన్సెప్ట్‌లు అవసరమైనప్పుడు. బలమైన ధృవీకరణ, పర్యవేక్షణ మరియు ఫీడ్‌బ్యాక్ లూప్‌ల కింద చిన్న భాగస్వాముల సమూహంతో GPT‑5.5‑Cyber మాకు నేర్చుకునే అవకాశాన్ని కల్పించే వర్క్‌ఫ్లోలు ఇవే.


గుర్తింపు మరియు పర్యవేక్షణ

బలహీనమైన సాఫ్ట్‌వేర్ ఇప్పటికే అమలులో ఉంటే, తదుపరి ప్రశ్న ఎవరైనా దాన్ని దుర్వినియోగం చేస్తున్నారా అనేది. EDR, SIEM, IGA/PAM మరియు మానిటరింగ్ భాగస్వాములు కొత్త అడ్వైజరీని ప్రత్యక్ష వాతావరణాల నుంచి వచ్చే సాక్ష్యాలుగా మారుస్తారు: టెలిమెట్రీ, అలర్ట్‌లు, డిటెక్షన్‌లు మరియు రెస్పాన్స్ వర్క్‌ఫ్లోలు. GPT‑5.5 విశ్లేషకులు ఆ సంకేతాలను అనుసంధానించడంలో, ముఖ్యమైన విషయాలను సంక్షిప్తంగా చెప్పడంలో, డిటెక్షన్‌ల ముసాయిదాలను రూపొందించడంలో, అలాగే డిస్క్లోజర్ నుంచి ఇన్వెస్టిగేషన్ దిశగా మరింత వేగంగా కదలడంలో సహాయపడగలదు. అదే చక్రం క్లౌడ్ వాతావరణాల్లో ప్రత్యేకంగా ముఖ్యమైనది, ఎందుకంటే అక్కడ ఎక్స్‌పోజర్, రిమీడియేషన్ మరియు డిటెక్షన్‌లు బలంగా అనుసంధానించబడి ఉంటాయి.

సాఫ్ట్‌వేర్ సరఫరా శృంఖల భద్రత

తదుపరి దశలో తెలిసిన హానికర కోడ్ ప్రొడక్షన్‌కు చేరకుండా ముందుగానే నిరోధించడం ఉంటుంది. ఒకసారి బలహీనత లేదా ప్యాకేజ్ కాంప్రమైజ్ అర్థమైతే, సాఫ్ట్‌వేర్ సప్లై చైన్ టూల్స్ ప్రమాదకర డిపెండెన్సీలు, దుష్ట అప్డేట్‌లు మరియు బలహీనమైన కోడ్ మార్గాలు కస్టమర్ వాతావరణాల్లో వ్యాప్తి చెందకముందే వాటిని ఆపడానికి సహాయపడతాయి. సైబర్ కోసం ట్రస్టెడ్ యాక్సెస్‌తో, GPT‑5.5 డిపెండెన్సీ మార్పులను పరిశీలించడం, స్వంత కోడ్‌లో ఎక్స్‌ప్లాయిటబిలిటీపై రీజనింగ్ చేయడం, రిమీడియేషన్‌కు ప్రాధాన్యత ఇవ్వడం, అలాగే డెవలప్‌మెంట్ చక్రంలోనే అనుమానాస్పద ప్యాకేజ్ ప్రవర్తనను ముందుగానే గుర్తించడంలో సహాయపడగలదు. 

Snyk, Gen Digital, Semgrep మరియు Socket వంటి భాగస్వాములు axios కాంప్రమైజ్ వంటి ఘటనలకు ఈ సామర్థ్యాలు ఎలా వర్తిస్తాయో మాకు పరీక్షించడంలో సహాయపడగలరు. అలాంటి సందర్భాల్లో అత్యంత వేగవంతమైన పరిష్కారం బలహీనమైన లేదా కాంప్రమైజ్ అయిన డిపెండెన్సీలు బిల్డ్‌లోకి ప్రవేశించకుండా నిరోధించడమే.

ఓపెన్ సోర్స్ మరియు రక్షకుల కోసం Codex Security

ఎకోసిస్టమ్ అంతటా బలహీనతలు వేగంగా వ్యాపించే మార్గాల్లో ఓపెన్ సోర్స్ ఒకటి. అందుకే మేము మెయింటైనర్‌లతో కలిసి అప్‌స్ట్రీమ్‌లో కూడా పెట్టుబడి పెడుతున్నాము. Codex Security కోడ్‌బేస్‌కు ప్రత్యేకమైన త్రెట్ మోడల్‌ను నిర్మించడం, వాస్తవిక దాడి మార్గాలను అన్వేషించడం, వేరుచేసిన వాతావరణాల్లో సమస్యలను ధృవీకరించడం, అలాగే మానవ సమీక్ష కోసం ప్యాచ్‌లను ప్రతిపాదించడం ద్వారా బృందాలు బలహీనతలను గుర్తించడానికి, ధృవీకరించడానికి మరియు పరిష్కరించడానికి సహాయపడుతుంది.

Codex for Open Source ద్వారా, కీలక ప్రాజెక్టుల ఎంపికైన మెయింటైనర్‌లు నిర్వహణ మరియు సమీక్ష భారం తగ్గించేందుకు Codex మరియు API క్రెడిట్‌లతో పాటు Codex Securityకు షరతులతో కూడిన యాక్సెస్‌ను పొందవచ్చు.

మేము Codex Security plugin ను విడుదల చేశాము. ఇది ప్రస్తుతం ఉన్న సెక్యూరిటీ వర్క్‌ఫ్లోను యాప్ లేదా CLI వంటి ఏ Codex ఇంటర్‌ఫేస్‌లోనైనా నేరుగా అందిస్తుంది, తద్వారా డెవలపర్లు త్రెట్ మోడలింగ్ నుంచి డిస్కవరీ, ధృవీకరణ, దాడి మార్గ విశ్లేషణ మరియు ధృవీకరించిన పరిష్కారాల దిశగా కదలగలరు.

ముందుకు చూస్తూ

మోడళ్లు సైబర్‌సెక్యూరిటీలో మరింత సామర్థ్యవంతంగా మారుతున్న కొద్దీ, ఆ సామర్థ్యాన్ని ఉత్తమంగా ఉపయోగించడం అంటే రక్షకులు బలహీనతలను వేగంగా కనుగొని సరిచేయడంలో సహాయపడటం. ఆ సామర్థ్యాలకు బాధ్యతాయుతంగా యాక్సెస్‌ను విస్తరించాలంటే, మోడల్‌ను ఎవరు ఉపయోగిస్తున్నారు, వారు ఏ వ్యవస్థలను లక్ష్యంగా తీసుకుంటున్నారు, మరియు ఆ పని అధీకృతమా అనే విషయాలపై మరింత బలమైన నమ్మకం అవసరం. గుర్తింపు మరియు సంస్థ ధృవీకరణ, ఆమోదిత వినియోగ పరిమితి, మరియు దుర్వినియోగ పర్యవేక్షణ మెరుగుపడుతున్న కొద్దీ, కాలక్రమేణా యాక్సెస్ మరింత విస్తరిస్తుందని మేము ఆశిస్తున్నాము.

Trusted Access for Cyberకు యాక్సెస్ పొందడం సులభం:

ఈ ప్రక్రియ ద్వారా ఆమోదం పొందిన అన్ని కస్టమర్‌లకు, ద్వంద్వ-ఉపయోగ సైబర్ కార్యకలాపాలపై ట్రిగ్గర్ కావచ్చిన రక్షణల చుట్టూ తక్కువ అడ్డంకులు కలిగిన ప్రస్తుత మోడళ్ల సంస్కరణలకు యాక్సెస్ లభిస్తుంది. దీంతో వారు సెక్యూరిటీ విద్య, రక్షణాత్మక ప్రోగ్రామింగ్, మరియు బాధ్యతాయుత బలహీనతల పరిశోధనను కొనసాగించగలరు. 

ఆల్ఫా పరీక్షల సమయంలో, GPT‑5.5‑Cyber ఇప్పటికే కీలక వ్యవస్థల ఆటోమేటెడ్ రెడ్-టీమింగ్‌ను విస్తరించడానికి మరియు అధిక తీవ్రత గల బలహీనతలను ధృవీకరించడానికి ఉపయోగించబడింది. బాధ్యతాయుత వెల్లడిలో భాగంగా దీన్ని భవిష్యత్తులో ఒక సాంకేతిక డీప్-డైవ్‌లో మేము వివరించబోతున్నాము.

సైబర్ కోసం ట్రస్టెడ్ యాక్సెస్‌ ద్వారా మా ఫ్లాగ్‌షిప్ మోడళ్లతో పాటు, GPT‑5.5‑Cyber వంటి ప్రత్యేక సైబర్ మోడళ్లు మరియు భవిష్యత్తులో మరింత సైబర్-సామర్థ్యవంతమైన మోడళ్లతోనూ రక్షకులను వేగవంతం చేయడం మేము కొనసాగిస్తామని భావిస్తున్నాము.

రచయిత

OpenAI