Orion-1 beats GPT-6 Sol in cyberattack benchmark, claims 80.8 attack reconstruction score

artemis security launched orion-1, an ai model designed to investigate and respond to cyberattacks on corporate networks. orion-1 connects log data from networks, endpoints, and cloud services to trace attacks back to a single actor and initiate countermeasures. on artemis' benchmark, orion-1 outperformed general-purpose models like claude opus 5.5 and gpt-6 sol, scoring 80.8 in attack reconstruction compared to 58.3 for gpt-6 sol. orion-1 is live in private preview for select customers on the artemis platform. this launch addresses challenges faced by security teams as ai guardrails meant to block attackers are now blocking defenders.