Experts Say There’s Now an Open Source AI Model as Scary as Mythos

Beijing-based AI company Z.ai has released GLM-5.2, an open-weight model that can perform large-scale coding tasks and identify vulnerabilities at a level comparable to Anthropic's Mythos. The model's weights are publicly available, allowing anyone to download and modify them.

Researchers have suggested that Z.ai may have distilled capabilities from OpenAI's GPT-5.5 and Anthropic's Opus 4.8 to achieve this performance. The model's release marks a significant milestone in open-source AI, as it approaches the capabilities of proprietary frontier systems.

The development has drawn attention from the AI research community, particularly regarding safety implications of openly available models with such advanced code-generation and security-assessment abilities. It also highlights the ongoing convergence between open-source and proprietary AI systems.