开源模型能力逼近前沿，安全差距依然显著
原标题：Open-weight AI models are catching up to the frontier. The safety gap remains.
AI 摘要
SaferAI 的报告显示，中国开源模型 GLM-5.2 在网络和生物能力上已接近 OpenAI 的 GPT-5.5 和 Anthropic 的 Claude Opus 4.7，但在安全实践上存在显著差距。GLM-5.2 未拒绝任何有害任务，而 Claude Opus 4.7 则一致拒绝。报告指出，开源模型一旦发布，其安全措施无法强制执行，而前沿开发者依赖分类器、拒绝训练等防护手段。SaferAI 建议通过预训练数据过滤等方法减少风险，但中国政策更关注政治敏感内容而非灾难性风险。
正文节选
As policymakers debate how to govern increasingly powerful AI systems like OpenAI’s GPT-5.6 Sol and Anthropic’s Mythos, a Chinese open-weight model has narrowed the gap with the industry’s leaders. GLM-5.2, the open-weight AI model from China’s Z.ai, is only a few months behind OpenAI’s GPT-5.5 and Anthropic’s Claude Opus 4.7 on cyber and bio capabilities, according to a new report from AI safety nonprofit SaferAI. But the divide between frontier capabilities and safety practices is growing. Acc