GLM-5.3来了，不重训基础模型，后训练拉高Terminal-Bench到28.3，代码和长程任务提升，两周后开放权重。
Z.ai 于 2026 年 8 月 14 日发布 GLM-5.3，复用 743B 参数的 GLM-5.2 基础模型。GLM-5.3 的全部增益来自扩展后训练，未改动基础模型权重。Terminal-Bench 3.0 得分从 4.6 升至 28.3，DeepSWE v1.1 从 46.2 升至 66.9。网络安全方面，CyberGym 达到 84.5%，ExploitBench 翻倍以上至 54.4%。GLM-5.3 的权重约两周后发布。
Z.ai Ships GLM-5.3 Without Retraining the Base Model: Better at Complex Coding and Long-Horizon Tasks
Z.ai released GLM-5.3 on August 14, 2026. The model reuses the 743B GLM-5.2 base unchanged. Every reported gain comes from scaled post-training: more long-horizon task environments, more environment types, longer training. Terminal-Bench 3.0 moves from 4.6 to 28.3, and DeepSWE v1.1 from 46.2 to 66.9. Cybersecurity moved further than Z.ai says it planned, with CyberGym at 84.5% and ExploitBench more than doubling to 54.4%. Weights arrive in about two weeks. The post Z.ai Ships GLM-5.3 Without Retraining the Base Model: Better at Complex Coding and Long-Horizon Tasks appeared first on MarkTechPost .