X @BSCN

UK AISI: GPT-5.5 MATCHES MYTHOS ON CYBER TASKSThe UK AI Security Institute (@AISecurityInst) found @OpenAI's GPT-5.5 matches or slightly edges Anthropic's Claude Mythos Preview on advanced cybersecurity benchmarks, per AISI's official evaluation published this week.GPT-5.5 also became the second model after Mythos Preview to complete AISI's 32-step "The Last Ones" corporate network attack simulation end-to-end, an exercise estimated to take a human expert around 20 hours. GPT-5.5 solved it in 2 of 10 attemp ...