AI offensive cyber capabilities are doubling every six months, safety researchers find
AI offensive cyber capabilities are rapidly advancing.
A recent study by Lyptus Research highlights a dramatic acceleration in AI's offensive cyber capabilities, which have been doubling every 5.7 months since 2024. This marks a significant increase from the previous rate of 9.8 months, with models like Opus 4.6 and GPT-5.3 Codex achieving a 50% success rate on complex tasks that would typically take human experts around three hours to complete. The findings underscore the rapid evolution of AI in cybersecurity, with open-source models trailing behind their closed-source counterparts by about 5.7 months.
Key Takeaways
- 1.
AI's offensive cyber capabilities have accelerated from doubling every 9.8 months to every 5.7 months since 2024.
- 2.
Models like Opus 4.6 and GPT-5.3 Codex can now complete tasks with a 50% success rate in about three hours.
- 3.
Open-source models lag behind closed-source counterparts by approximately 5.7 months.
Get your personalized feed
Trace groups the biggest stories, videos, and discussions into one feed so you can stay current without scanning ten tabs.
Try Trace free