日本語 
OrcaCyber, the cybersecurity team behind OrcaRouter, has launched an API version of Z.ai’s GLM-5.3 model with reduced rejection responses for security research. Available through Orca Cyber Tier, the model supports a context window of up to 1 million tokens and output limits of 128,000 tokens. It is designed to support security investigations, vulnerability testing, penetration testing, malware analysis and secure code reviews.
Also Read: Wiz Cloud Japan Unveils Cybersecurity Initiative for Japan
The offering follows OrcaCyber’s approach of using multiple AI models to identify infrastructure weaknesses before attackers can exploit them. Access requires an application and review, with the service intended for security researchers and red and blue teams. The launch also highlights a growing focus on specialized AI models for cybersecurity testing, alongside safeguards for production environments.
English