Z.ai’s GLM-5.2 Matches OpenAI’s GPT-5.5 on Enterprise Benchmarks

Z.ai’s GLM-5.2 has become the highest-ranked open-weight large language model, scoring within one percentage point of Anthropic’s Claude Opus 4.5 on ProofBench—the first open model to cross that threshold. With 750 billion parameters and a one million token context window, it signals that open-weight models are rapidly closing the performance gap with expensive proprietary alternatives.

CryptoBandits Malware Hid Undetected for Four Months Before Microsoft Disclosed It

CryptoBandits malware operated undetected for four months, using USB worms and Tor-based command channels to steal crypto keys and drain wallets before Microsoft’s June disclosure. The campaign targets seed phrases and private keys with precision, combining clipboard hijacking with self-replicating USB propagation to expand its attack surface beyond typical phishing threats.

Exit mobile version