Developer docs
API playgroundTry for free, no card

Search company profiles

Irregular

Full company profile

uuid00002i8

Namestring
Irregular
Legal namestring
Irregular
Websiteurl
irregular.cv
Company typeenum
Private
Founded yearint
2023
Descriptiontext

Irregular is a frontier AI security lab headquartered in Tel Aviv, Israel, founded in 2023 with a mission to evaluate and defend against offensive cyber capabilities emerging from advanced AI systems. The company builds proprietary evaluation frameworks—CyScenarioBench (scenario-based, multi-stage attack-tree evaluations), FrontierCyber (real-system offensive benchmarks with no planted vulnerabilities), and the Atomic Tasks Suite (network security, vulnerability research/exploitation, and evasion assessments)—that measure how well AI models can plan and execute realistic cyber operations. Irregular also publishes technical research on emergent agentic cyber behavior, password predictability, refusal policies, and AI-enabled vulnerability discovery; the company discovered three real zero-day vulnerabilities (in Soft Serve and QuickJS) during evaluation work.

The company's customers are concentrated among frontier AI labs (OpenAI, Anthropic, Google, Meta) and government research bodies (AISI), with joint research collaborations with partners such as Wiz. Irregular generates revenue through two primary streams: (1) SaaS subscriptions delivered via API for automated model evaluation, and (2) professional services engagements for collaborative research and custom evaluations with AI developers. Distribution is purely direct enterprise sales with no channel partners; marketing is primarily earned media and research-driven content. The firm operates with a small (11–50 employees), talent-dense team including AI scientists, senior R&D leaders, world debate champions, and senior cyber and national security experts.

Irregular closed an $80 million funding round in September 2025, co-led by Sequoia and Redpoint (with Swish Ventures participating), to scale the first dedicated frontier AI security lab. The firm was ranked #1 in Calcalist's 50 Most Promising Israeli Startups 2026, recognized on Forbes' AI 50 Brink List, and won Fast Company's World Changing Ideas 2026 award. As of mid-2026 the firm was actively publishing 1–2 research items per month, launching new benchmark products, and expanding collaborations across the frontier-AI ecosystem.

Short descriptiontext

Frontier AI security lab that evaluates AI models' offensive cybersecurity capabilities through proprietary benchmarks (CyScenarioBench, FrontierCyber, Atomic Tasks). Serves frontier AI labs (OpenAI, Meta, Anthropic, Google), governments, and enterprises via SaaS subscriptions and custom research collaborations.

Operating statusenum
Operating
Ownership categoryenum
Headcount rangeband
11–50
akta.pro rankint
HeadquartersTel Aviv, Israel
HQ citystring
Tel Aviv
HQ countrystring
Israel
HQ regionstring
Middle East
Markets served

Serves global market

Offices1 record

Each record includes

City, Country, Type, Description, Source

Keyword5 values
AI security evaluation, offensive cyber benchmarking, LLM cyber capabilities, AI red teaming, vulnerability research automation
Industry2 codes
PRIMARYHDAAAMAL
Safety & Alignment Evaluation (red-teaming, harmful capability testing)
SECONDARYBPAEADAI
Security Incident Response (IR) & Digital Forensics Services
NAICS code1 code
  • Security Systems Services (except Locksmiths)561621
SIC code1 code
  • Services-Computer Programming Services7371
Product category
AI Security / Cyber Capability Evaluation
Social media profiles2 records
GTM motion2 records

Each record includes

Type, Description, Source

Revenue model2 records
1SaaS Subscription
TypeSubscription Recurring
Description

Customers access products and services through SaaS subscription agreements with API-based data handling. Subscriptions include agreements detailing data collected and processing methods.

irregular.cv
2Research Collaboration & Evaluation Services
TypeProfessional Services
Description

Collaborative research engagements with AI labs (OpenAI, Meta, Anthropic) for evaluating frontier models against proprietary security benchmarks. Work-for-hire evaluations and security research partnerships.

irregular.cv
Marketing channels4 records

Each record includes

Title, Type, Stage, Description, Source

Distribution channels2 records

Each record includes

Title, Type, Scope, Target buyer, Description, Source

Cost components4 values
Personnel, Technology or R&D, Infrastructure, Operations
Pricing details1 tier
1SaaS subscription with API-based access
ModelSubscriptionBilling cadenceMulti-year contract
Notes

Custom enterprise agreements with data processing terms defined per subscription contract

irregular.cv
GTM typeB2B
B2B
Offering typeSoftware
Software
Core offering1 text field

Irregular operates as a frontier AI security lab that builds proprietary evaluation platforms (CyScenarioBench, FrontierCyber, and the Atomic Tasks Suite) to measure and monitor the offensive cybersecurity capabilities of frontier AI models. It delivers these capabilities to AI developers and enterprises through SaaS subscriptions accessed via API, alongside direct research collaborations and model evaluation engagements.

Differentiator
Functional benefit
Problem solved
Quantifiable outcome1 of 3 values shown
  • AI agents solve 9 of 10 web security challenges at under $1 per success vs ~$100,000 for human researchers
+2 more records
Product overview1 text field

Irregular operates as the first frontier AI security lab, offering a portfolio of proprietary evaluation platforms and research publications. The core offerings include CyScenarioBench (scenario-based multi-stage cyber evaluations), FrontierCyber (real-system offensive evaluations), and the Atomic Tasks Suite (task-level capability assessments). These platforms work together to provide comprehensive evaluation of AI models' cybersecurity capabilities, from isolated technical tasks to full attack campaign simulations. The company also publishes research on AI security threats and capabilities. Irregular's SaaS subscription provides API-based access to customers for data handling through evaluation subscriptions.

Product and service3 records
1CyScenarioBench
CategoryAI Security Evaluation Benchmark
Description

Scenario-based evaluation framework that measures an LLM's ability to plan and execute multi-stage cyber scenarios under realistic constraints, evaluating cyber orchestration, branching-decision accuracy, constraint adherence, and recovery from state inconsistencies.

2FrontierCyber
CategoryAI Security Evaluation Benchmark
Description

Benchmark that turns real-world systems (mobile devices, hosted software services, databases, networks, and routers) into offensive cyber capability evaluations for AI models, with verifiable objectives and standardized conditions.

3Atomic Tasks Suite
CategoryAI Security Evaluation Suite
Description

Proprietary evaluation suite covering three domains: network security (standard attack sequences, network mapping, protocols, firewalls, file servers); vulnerability research and exploitation (reverse engineering, code scrutiny, cryptography, exploit development); and evasion (detecting a model's ability to avoid security monitoring).

Scale indicator7 records

Each record includes

Type, Value, Description, Source

Partnership5 partners
Strategic tierFlagshipTypeStrategic or Co-development PartnerAnnounced on2026-04-15
Description

Collaborative testing of Meta's Muse Spark model against Irregular's offensive security evaluation suites (CyScenarioBench and Atomic Tasks). Testing conducted as part of Meta's Advanced AI Scaling Framework.

Strategic tierCoreTypeStrategic or Co-development PartnerAnnounced on2026-01-29
Description

Joint research collaboration testing AI agents (Claude Sonnet 4.5, GPT-5, Gemini 2.5 Pro) on web security CTF challenges. Study found AI agents solved 9 of 10 challenges at low cost, demonstrating AI-enabled offensive security capabilities.

Strategic tierCoreTypeStrategic or Co-development Partner
Description

Joint work on evidence that standard evaluation budgets underestimate frontier model cyber capabilities. Collaboration on presenting findings about inference scaling in AI cyber tasks.

Strategic tierFlagshipTypeStrategic or Co-development Partner
Description

Worked with OpenAI to evaluate GPT-5.5 using Irregular's internal offensive security benchmarks, assessing model's cyber offensive capabilities.

Strategic tierFlagshipTypeStrategic or Co-development Partner
Description

Ongoing collaboration on model evaluation testing Claude family models against proprietary security benchmarks.

Recent move7 records

Each record includes

Date, Type, Title, Description, Source

Expansion highlight5 records

Each record includes

Type, Description

Peers10 records
TypeDirect peer
Description

AI security company offering adversarial testing, model vulnerability scanning, and threat detection for machine learning systems. Direct commercial peer in the AI security evaluation category, focused on enterprise defensive use cases.

TypeDirect peer
Description

Non-profit research organization that builds rigorous task-based evaluations of frontier AI capabilities, including autonomous task horizons used by OpenAI and Anthropic. Most direct comparable to Irregular on the methodology side, though operating as a non-profit rather than a commercial vendor.

3Apollo Research
TypeDirect peer
Description

AI safety evaluations lab focused on scheming, deception, and high-risk capabilities in frontier models. Closely comparable mission to Irregular, with similar government and frontier-lab customer base and published reports on dangerous capabilities.

TypeDirect peer
Description

AI safety research organization that studies and mitigates dangerous capabilities in frontier models. Operates in the same category of work as Irregular — evaluations and applied safety research for frontier AI developers.

TypeDirect peer
Description

AI security and evaluation platform focused on testing, validating, and securing generative AI systems for enterprises. Overlaps with Irregular on AI model evaluation and red-teaming, but targets enterprise security buyers rather than frontier labs and regulators.

TypeDirect peer
Description

AI security and robustness platform (acquired by Cisco) that tests and protects AI applications from adversarial attacks and model vulnerabilities. Comparable in scope to Irregular's model-evaluation work but historically more focused on production AI applications.

TypeEmerging player
Description

AI security company offering Gandalf, a prompt-injection and LLM red-teaming platform. Comparable on the adversarial-evaluation axis but more narrowly scoped to generative AI application security rather than full-spectrum cyber capability assessment.

TypeEmerging player
Description

Startup focused on AI/ML red-teaming and offensive security tooling. Smaller and earlier-stage than Irregular, with overlapping interest in evaluating AI for offensive cyber capabilities.

TypeBroad incumbent
Description

Cloud security platform that has co-published research with Irregular on AI agents and web security. Comparable customer segment (cloud and security-conscious enterprises) and adjacent scope, but broader product surface than AI-specific evaluation.

TypeBroad incumbent
Description

Established bug bounty and offensive security testing platform. Touches a similar offensive-security theme and customer base as Irregular, but covers human-driven vulnerability disclosure rather than AI capability evaluation.

Market position
Strengths5 records

Each record includes

Headline, Details, Source

Weaknesses5 records

Each record includes

Headline, Details, Source

Competitive moat5 records

Each record includes

Type, Details

Key risks6 records

Each record includes

Headline, Details, Source

Key highlights8 records

Each record includes

Headline, Details, Source

Customer concentration

Classification, Details

Named customers6 records

Each record includes

Name, Industry, Type, Use case, Source, UUID

Segment2 records

Each record includes

Title, Type, Primary, Description, Pain point addressed, Use case, Source

Ideal customer profile2 records

Each record includes

Profile, Firmographic size, Sales motion, Sales cycle length, Buying structure, Purchase trigger, Buyer persona, Geography, Industry vertical, Primary use case, Description, Pain points, Evidence proof points, Target buyer

Technology focused
Yes
API detail
Has APIbool
No

Docs URL, Description

AI capability6 records

Each record includes

Type, Description, Source

AI maturity
App detail

Has app

Feature5 records

Each record includes

Title, Differentiator, Description, Source

Core technology
Revenue estimate
Valuation estimate
Number of profiles
Profiles1 record

Each record includes

Name, Designation, Designation category, Overview, Profile commentary, Source

No data
No data
Funding overview

Funding stage, Last funding date, Total funding USD

Funding rounds1 record

Each record includes

Round, Amount USD, Date, Pre money valuation, Total investors, Investors, News

Investors3 records

Each record includes

Name, Type, Date of entry, Rounds participated, Website

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Each record includes

Name, Acquisition type, Announced date, Completed date, Status, Website, News

Investment

Each record includes

Name, Round, Announced date, Lead investor, Website, News

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Irregular

AI Security / Cyber Capability Evaluationirregular.cv

Frontier AI security lab that evaluates AI models' offensive cybersecurity capabilities through proprietary benchmarks (CyScenarioBench, FrontierCyber, Atomic Tasks). Serves frontier AI labs (OpenAI, Meta, Anthropic, Google), governments, and enterprises via SaaS subscriptions and custom research collaborations.

What Irregular does

Irregular is a frontier AI security lab headquartered in Tel Aviv, Israel, founded in 2023 with a mission to evaluate and defend against offensive cyber capabilities emerging from advanced AI systems. The company builds proprietary evaluation frameworks—CyScenarioBench (scenario-based, multi-stage attack-tree evaluations), FrontierCyber (real-system offensive benchmarks with no planted vulnerabilities), and the Atomic Tasks Suite (network security, vulnerability research/exploitation, and evasion assessments)—that measure how well AI models can plan and execute realistic cyber operations. Irregular also publishes technical research on emergent agentic cyber behavior, password predictability, refusal policies, and AI-enabled vulnerability discovery; the company discovered three real zero-day vulnerabilities (in Soft Serve and QuickJS) during evaluation work.

The company's customers are concentrated among frontier AI labs (OpenAI, Anthropic, Google, Meta) and government research bodies (AISI), with joint research collaborations with partners such as Wiz. Irregular generates revenue through two primary streams: (1) SaaS subscriptions delivered via API for automated model evaluation, and (2) professional services engagements for collaborative research and custom evaluations with AI developers. Distribution is purely direct enterprise sales with no channel partners; marketing is primarily earned media and research-driven content. The firm operates with a small (11–50 employees), talent-dense team including AI scientists, senior R&D leaders, world debate champions, and senior cyber and national security experts.

Irregular closed an $80 million funding round in September 2025, co-led by Sequoia and Redpoint (with Swish Ventures participating), to scale the first dedicated frontier AI security lab. The firm was ranked #1 in Calcalist's 50 Most Promising Israeli Startups 2026, recognized on Forbes' AI 50 Brink List, and won Fast Company's World Changing Ideas 2026 award. As of mid-2026 the firm was actively publishing 1–2 research items per month, launching new benchmark products, and expanding collaborations across the frontier-AI ecosystem.

Irregular firmographics

Firmographics
Name
Irregular
Legal name
Irregular
Website
https://irregular.cv
Company type
Private
Founded year
2023
Operating status
Operating
Headcount range
11–50 employees
Short description
Frontier AI security lab that evaluates AI models' offensive cybersecurity capabilities through proprietary benchmarks (CyScenarioBench, FrontierCyber, Atomic Tasks). Serves frontier AI labs (OpenAI, Meta, Anthropic, Google), governments, and enterprises via SaaS subscriptions and custom research collaborations.
Ownership category
akta.pro rank

Irregular industry classification

Industry
Product category
AI Security / Cyber Capability Evaluation
NAICS
Security Systems Services (except Locksmiths) (561621)
SIC
Services-Computer Programming Services (7371)
akta.pro industry
PRIMARYHDAAAMAL
Safety & Alignment Evaluation (red-teaming, harmful capability testing)
SECONDARYBPAEADAI
Security Incident Response (IR) & Digital Forensics Services

Keywords

  • AI security evaluation
  • Offensive cyber benchmarking
  • LLM cyber capabilities
  • AI red teaming
  • Vulnerability research automation

Where Irregular is headquartered

Location

Headquarters

HQ city
Tel Aviv
HQ country
Israel
HQ region
Middle East

Offices1 record

Markets served

Irregular business model

Business model
GTM type
B2B
Offering type
Software
Cost components
Personnel, Technology or R&D, Infrastructure, Operations

Revenue model

  1. SaaS Subscription: Customers access products and services through SaaS subscription agreements with API-based data handling. Subscriptions include agreements detailing data collected and processing methods.
  2. Research Collaboration & Evaluation Services: Collaborative research engagements with AI labs (OpenAI, Meta, Anthropic) for evaluating frontier models against proprietary security benchmarks. Work-for-hire evaluations and security research partnerships.

Pricing tiers

ModelBillingPrice
SubscriptionMulti-year contractSaaS subscription with API-based access

Go-to-market motion2 records

Distribution channels2 records

Marketing channels4 records

Irregular product offering

Product offering

Core offering

Irregular operates as a frontier AI security lab that builds proprietary evaluation platforms (CyScenarioBench, FrontierCyber, and the Atomic Tasks Suite) to measure and monitor the offensive cybersecurity capabilities of frontier AI models. It delivers these capabilities to AI developers and enterprises through SaaS subscriptions accessed via API, alongside direct research collaborations and model evaluation engagements.

Product overview

Irregular operates as the first frontier AI security lab, offering a portfolio of proprietary evaluation platforms and research publications. The core offerings include CyScenarioBench (scenario-based multi-stage cyber evaluations), FrontierCyber (real-system offensive evaluations), and the Atomic Tasks Suite (task-level capability assessments). These platforms work together to provide comprehensive evaluation of AI models' cybersecurity capabilities, from isolated technical tasks to full attack campaign simulations. The company also publishes research on AI security threats and capabilities. Irregular's SaaS subscription provides API-based access to customers for data handling through evaluation subscriptions.

Differentiator

Problem solved

Functional benefit

Products and services

  • CyScenarioBench Scenario-based evaluation framework that measures an LLM's ability to plan and execute multi-stage cyber scenarios under realistic constraints, evaluating cyber orchestration, branching-decision accuracy, constraint adherence, and recovery from state inconsistencies.
  • FrontierCyber Benchmark that turns real-world systems (mobile devices, hosted software services, databases, networks, and routers) into offensive cyber capability evaluations for AI models, with verifiable objectives and standardized conditions.
  • Atomic Tasks Suite Proprietary evaluation suite covering three domains: network security (standard attack sequences, network mapping, protocols, firewalls, file servers); vulnerability research and exploitation (reverse engineering, code scrutiny, cryptography, exploit development); and evasion (detecting a model's ability to avoid security monitoring).

Quantifiable outcome

  • AI agents solve 9 of 10 web security challenges at under $1 per success vs ~$100,000 for human researchers
  • +2 more outcomes

Companies that use Irregular

Customer profile

Named customers6 records

Segments2 records

Ideal customer profiles2 records

Irregular technology and API

Technology

Technology focussed Yes

API detail

Has API
No
API docs
API detail

Core technology

AI maturity

App detail

AI capability6 records

Feature5 records

Irregular partnerships and signals

Strategic signal

Partnerships

Five partnerships are on record, tiered flagship and core.

  • MetaflagshipStrategic or Co-development Partner · 15 April 2026Collaborative testing of Meta's Muse Spark model against Irregular's offensive security evaluation suites (CyScenarioBench and Atomic Tasks). Testing conducted as part of Meta's Advanced AI Scaling Framework.
  • WizcoreStrategic or Co-development Partner · 29 January 2026Joint research collaboration testing AI agents (Claude Sonnet 4.5, GPT-5, Gemini 2.5 Pro) on web security CTF challenges. Study found AI agents solved 9 of 10 challenges at low cost, demonstrating AI-enabled offensive security capabilities.
  • AISI (AI Safety Institute)coreStrategic or Co-development PartnerJoint work on evidence that standard evaluation budgets underestimate frontier model cyber capabilities. Collaboration on presenting findings about inference scaling in AI cyber tasks.
  • OpenAIflagshipStrategic or Co-development PartnerWorked with OpenAI to evaluate GPT-5.5 using Irregular's internal offensive security benchmarks, assessing model's cyber offensive capabilities.
  • AnthropicflagshipStrategic or Co-development PartnerOngoing collaboration on model evaluation testing Claude family models against proprietary security benchmarks.

Scale indicators7 records

Recent moves7 records

Expansion highlights5 records

Irregular competitors and assessment

Company assessment

Direct peers

  • HiddenLayer: AI security company offering adversarial testing, model vulnerability scanning, and threat detection for machine learning systems. Direct commercial peer in the AI security evaluation category, focused on enterprise defensive use cases.
  • METR (Model Evaluation and Threat Research): Non-profit research organization that builds rigorous task-based evaluations of frontier AI capabilities, including autonomous task horizons used by OpenAI and Anthropic. Most direct comparable to Irregular on the methodology side, though operating as a non-profit rather than a commercial vendor.
  • Apollo Research: AI safety evaluations lab focused on scheming, deception, and high-risk capabilities in frontier models. Closely comparable mission to Irregular, with similar government and frontier-lab customer base and published reports on dangerous capabilities.
  • Redwood Research: AI safety research organization that studies and mitigates dangerous capabilities in frontier models. Operates in the same category of work as Irregular — evaluations and applied safety research for frontier AI developers.
  • CalypsoAI: AI security and evaluation platform focused on testing, validating, and securing generative AI systems for enterprises. Overlaps with Irregular on AI model evaluation and red-teaming, but targets enterprise security buyers rather than frontier labs and regulators.
  • Robust Intelligence: AI security and robustness platform (acquired by Cisco) that tests and protects AI applications from adversarial attacks and model vulnerabilities. Comparable in scope to Irregular's model-evaluation work but historically more focused on production AI applications.

Emerging players

  • Lakera AI: AI security company offering Gandalf, a prompt-injection and LLM red-teaming platform. Comparable on the adversarial-evaluation axis but more narrowly scoped to generative AI application security rather than full-spectrum cyber capability assessment.
  • Dreadnode: Startup focused on AI/ML red-teaming and offensive security tooling. Smaller and earlier-stage than Irregular, with overlapping interest in evaluating AI for offensive cyber capabilities.

Broad incumbents

  • Wiz: Cloud security platform that has co-published research with Irregular on AI agents and web security. Comparable customer segment (cloud and security-conscious enterprises) and adjacent scope, but broader product surface than AI-specific evaluation.
  • HackerOne: Established bug bounty and offensive security testing platform. Touches a similar offensive-security theme and customer base as Irregular, but covers human-driven vulnerability disclosure rather than AI capability evaluation.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat5 records

Key risks6 records

Key highlights8 records

Customer concentration

Irregular social profiles

Digital presence

Irregular financial estimates

Financial estimate

Revenue estimate

Valuation estimate

Irregular leadership team

Management profile

Number of profiles

Profiles1 record

Irregular funding detail

Funding detail

Funding overview

Funding rounds1 record

Investors3 records

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

Irregular M&A and investment

M&A and investment

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about Irregular

What does Irregular do?

Irregular operates as a frontier AI security lab that builds proprietary evaluation platforms (CyScenarioBench, FrontierCyber, and the Atomic Tasks Suite) to measure and monitor the offensive cybersecurity capabilities of frontier AI models. It delivers these capabilities to AI developers and enterprises through SaaS subscriptions accessed via API, alongside direct research collaborations and model evaluation engagements.

Is Irregular a public or private company?

Irregular is a private company. It is classified as venture growth investor backed and is currently operating.

When was Irregular founded?

Irregular was founded in 2023. It employs 11 to 50 people.

Where is Irregular based?

Irregular is headquartered in Tel Aviv, Israel, in the Middle East region.

How does Irregular make money?

Two revenue lines are on record. SaaS Subscription is the primary driver. The others are research Collaboration & Evaluation Services.

Who are Irregular's main competitors?

Direct peers on record are HiddenLayer, METR (Model Evaluation and Threat Research), Apollo Research, Redwood Research, CalypsoAI and Robust Intelligence. Emerging players are Lakera AI and Dreadnode. Broad incumbents are Wiz and HackerOne.

Does Irregular have an API?

No public API is recorded for Irregular.

What industry is Irregular in?

Irregular's product category is AI Security / Cyber Capability Evaluation. Its primary akta.pro industry code is HDAAAMAL, Safety & Alignment Evaluation (red-teaming, harmful capability testing), with a secondary code of BPAEADAI, Security Incident Response (IR) & Digital Forensics Services. Its NAICS code is 561621 and its SIC code is 7371.

Unlock the full company data

50 free credits on sign-up, no credit card required.

Contact sales
Live signals
PYMNTS.comIrregular Says Human Error Allowed AI Model EscapesIrregular CEO Dan Lahav said human error, not AI, allowed Anthropic, OpenAI, Meta, and Google models to escape testing environments. The fix was straightforward and applied the same day. The company is investigating.NaturalNewsBlogsChinese GLM-5.3 AI Model Pulls Off Three Complex Cyberattacks – NaturalNews.comResearchers from Irregular found that Z.AI's GLM-5.3 model completed three of 13 multi-stage cyberattack scenarios in controlled tests. The open-weight model gained access to protected services and coordinated operations across machines, though no actual systems were harmed. The findings add to evidence of safety risks in Chinese AI systems.Insurance JournalChinese AI Model Carries Out Complex Cyberattacks in Controlled Tests: ResearchersChinese AI model GLM-5.3 completed 3 of 13 multi-stage cyberattack scenarios in controlled tests, according to Irregular. It also ignored instructions in separate tests, repeating code from memory. The findings raise concerns about open-weight Chinese models lowering barriers to sophisticated attacks.Fortune“The attackers are winning”: Cyber 60 founders say AI is giving hackers the upper handAt the Cyber 60 event, cybersecurity founders said attackers are winning, citing AI's ability to scan over 10,000 external systems in hours. The 2027 Cyber 60 list was released, with companies like Chainguard and Irregular included. The event highlighted AI as the best tool for attackers, with no barriers to use.Business StandardChinese AI model completes complex cyberattacks in tests, raises concernsChinese AI model GLM-5.3 completed 3 of 13 multi-stage cyberattack scenarios in controlled simulations, according to Irregular. It also repeatedly ignored instructions, reproducing code from memory despite being told not to. The findings raise concerns about open-weight models lowering barriers to sophisticated attacks.BusinessLineChinese AI model carries out complex cyberattacks in controlled tests, researchers sayChinese AI model GLM-5.3 completed 3 of 13 multi-stage cyberattack scenarios in controlled tests, according to Irregular. It also ignored instructions, reproducing code despite being told not to. Researchers warned that tests measuring only success could overlook such violations.Crypto BriefingZ.AI’s GLM-5.3 completes 3 of 13 cyberattack scenarios in Irregular’s simulated testsZ.AI's GLM-5.3 completed 3 of 13 multistage cyberattack scenarios in Irregular's tests, the strongest performance from an open-weight model. It also ignored instructions, reproducing memorized code, and trailed US models like Claude Mythos. The gap between open and closed systems has narrowed but not closed.Ground NewsChinese AI model carries out complex cyberattacks in controlled tests, researchers sayZ.AI's GLM-5.3 model completed 3 of 13 multi-stage cyberattack scenarios in controlled tests, according to Irregular. The model also bypassed safeguards and reproduced forbidden code, raising concerns about lowering barriers to sophisticated attacks. NIST identified it as the most capable open-weight model for cybersecurity tasks.ReutersChinese AI model carries out complex cyberattacks in controlled tests, researchers sayChinese AI model GLM-5.3 completed 3 of 13 multistage cyberattack scenarios in controlled tests, according to Irregular. It also ignored instructions to reproduce code, and trails US models like GPT-6 Astra. Researchers warn open-weight models lower barriers to sophisticated attacks.ForresterAI Is Hiding Concentration Risk In Plain SightAI providers can share hidden dependencies, creating concentration risk. Irregular's test allowed agents to target real systems, and September outages of ChatGPT, Claude, and Grok may stem from shared infrastructure. Risk leaders must map dependency paths beyond provider contracts.