Z.ai的GLM-5.3在安全测试基准上超越专有模型，仅靠微调实现，还因能力过强暂不公开权重。
Z.ai的GLM-5.3模型在CyberGym漏洞基准测试中达到84.5%的准确率。这一成绩超越了多个顶级专有模型。相比前代GLM-5.2，GLM-5.3实现了显著性能提升。Z.ai团队通过微调和优化模型智能体能力实现这一突破，未更改基础模型。
💻 Z .ai's GLM-5.3 just hit 84.5% on the CyberGym vulnerability benchmark, beating top proprietary m...
💻 Z .ai's GLM-5.3 just hit 84.5% on the CyberGym vulnerability benchmark, beating top proprietary models, a huge gain over the performance of its predecessor GLM-5.2. The kicker? Z.ai d’s AI engineers did it purely through fine-tuning and optimization of the model’s agentic capabilities, without changing the base model. The model grew so capable at finding and targeting potential exploits that Z.ai d held back the open weights for safety testing. Read the full analysis in The Batch: hubs.la/Q04w3GkF0 A 📖 #DeepLearningAI A #Cybersecurity t #LLMs Ms 💬 2 🔄 0 ❤️ 8 👀 1520 📊 3 ⚡