Google wants to compete with Claude Mythos with its new CodeMender, here is how
Follow UsGoogle expanded access to its AI security tool CodeMender at Google I/O 2026.
CodeMender can find software flaws and suggest fixes before issues grow.
Google joins Anthropic and OpenAI in the race for AI cybersecurity tools.
Google is stepping deeper into the cybersecurity race as major AI companies look for new ways to grow beyond chatbots and content tools. The company announced wider access to CodeMender, an AI-powered system designed to detect and fix weaknesses in software code during its Google I/O 2026 event. The tool was first introduced quietly last year. However, Google is now presenting it as a serious product for governments and businesses. The move comes at a time when artificial intelligence firms are competing to prove their technology can handle high-level security work. With cyberattacks becoming more common and expensive, AI companies see security tools as both a business opportunity and a way to show the practical value of their most advanced models today.
SurveyGoogle is paying more attention again after Anthropic announced its new Claude Mythos Preview model. People got very interested in it because some reports said it was so powerful that it could not be fully released to the public. This news started many discussions in the tech world, including with banks and government offices that want better digital security systems.
Also read: Google I/O 2026: Samsung and Google unveil Gemini-powered smart glasses to rival Ray-Ban Meta
Soon after Anthropic’s reveal, OpenAI introduced its own security-focused AI offering. Google has now joined that growing competition with CodeMender. According to Google DeepMind Chief Technology Officer Koray Kavukcuoglu, the company wants the tool to help secure software systems by not only identifying vulnerabilities but also suggesting fixes before problems become larger threats.
Kavukcuoglu said Google has already started discussions with enterprise customers and government groups interested in using the system to review and audit important digital infrastructure. The company believes AI models trained on large amounts of code can speed up security checks that often take human teams weeks or months to complete.
Google CEO Sundar Pichai also acknowledged the growing importance of large AI models in cybersecurity. During a press briefing, he credited Anthropic for showing the value of such systems in security-related tasks, while adding that Google has the capability to compete in the same space.
