-
Tweets140
-
Followers1K
-
Following1
-
Likes12
Read on our website: irregular.com/research/asses…
GPT-5.6 Sol demonstrated capability slightly stronger than GPT-5.5. It discovered vulnerabilities more consistently than it could compose them into reliable attack paths under production defenses, with clear limitations against hardened targets and over long horizons.
We worked with @OpenAI to evaluate GPT-5.6 Sol, including the first deployment of FrontierCyber as part of a frontier model assessment with a partner. FrontierCyber measures offensive-cyber capability on real, off-the-shelf systems, with no planted vulnerabilities and no predefined exploit paths. The model is not told where to look or how to attack.
Read on our blog: irregular.com/research/front…
Initial evaluations are already surfacing previously unknown vulnerabilities, now moving through responsible disclosure. For example, a model built a novel multi-vulnerability chain to gain unauthorized access to private information on a widely used mobile device.
Introducing the FrontierCyber benchmark: Irregular’s new approach to advanced offensive-cyber evaluations. It measures AI models’ offensive skills on real systems, including mobile devices, hosted software services, databases, and networks.
At @ManGroup's Technology Offsite this week, our CEO @dan_lahav gave the keynote on frontier AI security risk as a category of its own, alongside classical cybersecurity. The tools we defend networks with were built for systems that follow rules. AI systems reason toward a goal, and when a rule sits in the way, working around it is in scope. This is an emerging class of risk: a capable model inside your environment, reasoning faster than any person and in ways that aren't fully transparent, toward objectives that may not match yours. The hard part isn't that models are malicious, it's that they're effective. Thanks to Man Group for having Dan and for the conversation. These are the questions enterprises are starting to take seriously, and we're focused on shaping the answers.
Full Fast Company writeup: fastcompany.com/91547245/pushi… The World Changing Ideas list: fastcompany.com/world-changing…
Most 'world-changing' AI ideas are about what these systems can do. Ours is about whether you can trust them to do it. We’re proud to be a winner on @FastCompany’s 2026 World Changing Ideas list, alongside the labs and teams betting on getting this right.
More about Irregular's CyScenarioBench here: irregular.com/research/cysce…
We’re happy to share that CyScenarioBench, our benchmark for offensive cyber operations, was used by @AnthropicAI to test Claude Mythos 5 and Claude Fable 5. Most current cybersecurity evaluations check isolated skills, such as vulnerability research or exploitation. CyScenarioBench measures a more complex aspect of cybersecurity: whether an AI system can plan and execute a full attack across multiple stages in a realistic environment. As offensive capabilities advance, CyScenarioBench is among the few benchmarks that are not saturated and help differentiate between model capabilities.
Read the full writeup here: nytimes.com/2026/06/02/tec…
The New York Times covered new research from the University of Toronto on AI-powered worms. Speaking to @nytimes, our CEO @dan_lahav highlighted the gap between lab demonstrations and real-world cyber impact: reliability, complexity, and defenses. At Irregular, we work on widening that gap so defenders can move faster as AI capabilities advance.
Honored to be the main sponsor of CyberML 2026, a leading technical conference dedicated to the intersection of cybersecurity and machine learning. Our co-founder and CTO, Omer Nevo, opens with the keynote "Artificial Attackers: Risks, Capabilities and Mitigations.” Swing by our booth. We’re hiring AI/ML researchers, cyber researchers & research engineers. Link and tickets below 👇
Full list and report: redpoint.com/infrared/repor…
Thrilled to be recognized in @Redpoint's 2026 InfraRed 100, highlighting 100 of the most promising private companies in AI infrastructure. This recognition is a powerful validation of our mission: to protect the world as AI systems become increasingly capable and sophisticated.
𝔼 @eMuyi
422 Followers 2K Following
Nadav Kahlon @nadav_kahlon
2 Followers 26 Following
Daniel Christa @Auxbilb
2 Followers 108 Following
AI for Thinking @AIforThinking
16 Followers 1K Following
Thomas Houhou @Th0h0
422 Followers 1K Following Founder of DeepProof (AI for Formal Verification) | Security researcher - found vulnerabilities in @chainlink, @OpenAI, @AnthropicAI, @X | Cybersecurity @ETH
odyotertoc @odyotertoc
1 Followers 19 Following
何 @AvimanyuRoy3
755 Followers 6K Following
Svitlana Vakulenko �... @svakulenk0
1K Followers 1K Following Assistant Professor (tenure track) @wu_vienna #convsearch Former @Apple @AmazonScience PhD @tu_wien https://t.co/FFUu6h0cAc
Gabriele Sarti @gsarti_
2K Followers 2K Following Open-source interpretability to seize the means of prediction. Postdoc w/ @davidbau @ndif_team @Northeastern. Prev: @GroNLP, @amazonscience
Kerem Zaman ✈️ AC... @KeremZaman3
978 Followers 2K Following student researcher @GoogleDeepMind and PhD student @unc_ai_group working on explainable, interpretable and safe AI systems | prev. BSc @UniBogazici | kal '18
Hussein Muhaisen @husseinmuhaisen
2K Followers 4K Following Computer Security Researcher @OrbitCurveSec, @(PagedOut_zine), @(GuidedHacking).
Ram Keidar @RamKeidar1
15 Followers 337 Following
Wanderer @Wanderer_hl
184 Followers 864 Following
???Driver??? @DriveIntel
35 Followers 2K Following
Gal Ringel @GalRingel
831 Followers 1K Following CSO at @sayminetech. Dad (+2), Tech geek, an entrepreneur by heart, and an ex-VC investor. @forbes under30. #privacy #security
Asaf Yitzhaki @YitzhakiAsaf
22 Followers 774 Following
Chaim @haim770
126 Followers 3K Following
Omri Segev Moyal @GelosSnake
10K Followers 1K Following I used to take things apart, now I build them | CEO @proferosec | @forbes 30 under 30. Co-founder @minervalabs (Acquired by @rapid7)
Talten @Taltenisbor
2 Followers 505 Following
S. Daniel Leon @sdanielleon
346 Followers 1K Following Rose fast. Fell hard. Building again. CEO @ K Group. Chairman @ MAD Growth. AI, Investing, Geopolitics & Israel.
ido @Ido36236332
0 Followers 362 Following
Idan G @IdanDG
0 Followers 171 Following
DC @DexterCasey
14 Followers 324 Following
Kaelan @kaelanyim
52 Followers 663 Following roboticist | prev. AI safety, RL for missiles | cave diver
הלל דותן @HallelDotan
2 Followers 182 Following
Deep Dive R&D @MetorOkSo
5 Followers 143 Following 8 years multidisciplinary R&D (hardware, not software!). Building complex systems. Dive instructor. Parkour. 🤿🛠️
Na @NaLmnop
47 Followers 452 Following
Roei_d @d_roei
10 Followers 1K Following
Slashy @Slashy_fx
146 Followers 88 Following Founder of @Protocol01_ Building the privacy SDK for Solana Any app can add private transactions in 10 lines of code. Rust · TypeScript · ZK-SNARKs · STARKs
Sharon Besser @sbesser
811 Followers 1K Following Tinker, Tailor, Fixer (not a spy). Cyber, thought loser, 4 kids with true love. vi/vim. product @ Sepio I'll pay for the ☕
Steph @stephmor1807
0 Followers 19 Following
Thanos Tarasidis @ATarasidiss
604 Followers 3K Following Tech-savvy dad, trailblazer in workflow automation, and dedicated to empowering businesses to unlock value through the ServiceNow platform.
Eden Marco @EdenEmarco177
4K Followers 516 Following LLMs @google cloud | Best-seller @udemy Instructor | Opinions stated here are my own, not those of my company
Clara Suslik @ClaraSuslik
39 Followers 132 Following
Doron Adler @Norod78
4K Followers 3K Following Fooling around with Generative machine learning models since 2019
אחשליץ האדו... @AchShelitz
526 Followers 289 Following הייטקיסט | אוהד הפועל באר שבע רוב שעות היממה | 'שמאלני בוגד'
Iron-Side1904 @ShafirMatthew
63 Followers 1K Following
Yevgeni Nogin @yevgenin
41 Followers 645 Following
Alcides Fonseca @alcides
2K Followers 3K Following Typing types. Associate Professor at the University of Lisbon.
Ben @Benmam97
36 Followers 2K Following
Omer Adler @adler_omer
49 Followers 1K Following
Avihay Bar @AvihayBar
256 Followers 3K Following Software, Tech & Computer Graphics addict. Fluent in Python, Hand-waving and Shaders. Great people skills and a poor sense of humor. Views are my own.
You might like
