GLM-5.3 发布：中国实验室如何保持前沿竞争力
原标题：GLM-5.3: How Chinese labs keep stride with the frontier
AI 摘要
Z.ai 发布了 GLM-5.3 模型，目前仅在编程计划中提供，API 即将上线，两周后开放权重。该模型在多个基准测试中超越了 Moonshot AI 的 Kimi K3，部分指标甚至超过 Claude Fable 5 和 GPT-5.6-Sol，且参数量仅为 Kimi K3 的三分之一。Z.ai 表示 GLM-5.3 基于 GLM-5.2 的相同基础模型，仅通过扩展后训练实现性能提升。分析认为，中国实验室能保持前沿地位的关键在于更快的发布周期和更强的后训练能力，而非简单的蒸馏。
正文节选
GLM-5.3: How Chinese labs keep stride with the frontier Hint: It’s really not a distillation story. Housekeeping: I’m traveling so cannot make a voiceover for this post. Today, Z.ai announced their GLM-5.3 model, currently only available in the coding plan, coming soon to their API and in two weeks’ time to Hugging Face (open weights). This model looks exceptional, with a somewhat astounding increase in scores. On many benchmarks the model has surpassed Moonshot AI’s Kimi K3 and on some it’s sur