Wiz, now part of Google, has launched Project Atlas, an AI system designed for vulnerability research that has outperformed leading models like Anthropic's Mythos and OpenAI's GPT-5.5 Cyber on the CyberGym benchmark. Unlike single-model approaches, Atlas utilizes a network of specialized AI agents that collaborate to select the most effective model for each task, including models like ChatGPT and Claude Opus. This system has demonstrated a lower rate of hallucinations and false positives, and has already identified over 200 previously unknown vulnerabilities, including a critical flaw in GitHub.com. AI
IMPACT This system's success suggests a shift towards multi-agent AI architectures for complex tasks like vulnerability discovery, potentially accelerating defensive cybersecurity capabilities.
RANK_REASON Product launch of a novel AI system for cybersecurity that demonstrates superior performance on a key benchmark. [lever_c_demoted from significant: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
