Zhizhi released GLM-5.3, based on the same foundation as GLM-5.2, enhancing intelligence through extreme post-training scaling. Programming ability increas...
Zhizhi released GLM-5.3, based on the same foundation as GLM-5.2, enhancing intelligence through
extreme post-training scaling. Programming ability increased by 50% compared with the previous generation, ranking first among open-source models in publicly available benchmarks like Terminal Bench 3.0, and approaching Claude Fable 5. The model performs on par with Mythos 5 in white-box code review and other security tasks, scoring 84.5% in the CyberGym test. Model weights will be open-sourced in two weeks, and tools such as ZCode and AutoClaw are now online. 🔗 Read original via AIHOT · https://aihot.virxact.com/items/cmssir12d047oroffnpn5pcx1
智谱发布GLM-5.3，基于与GLM-5.2相同的基座，通过极致的后训练Scaling提升智能上界，编程能力较前代提升50%，在Terminal Bench
3.0等公开基准中取得开源第一，并接近Claude Fable 5。
模型在白盒代码审查等安全任务中表现持平Mythos 5，在CyberGym测试中得分84.5%。
模型权重将在两周后开源，即日起上线ZCode、AutoClaw等工具。
🔗 阅读原文 via AIHOT · https://aihot.virxact.com/items/cmssir12d047oroffnpn5pcx1
智谱发布GLM-5.3，称其与GLM-5.2采用相同基座，并通过后训练Scaling提升能力。材料称其编程能力较前代提升50%，在部分公开基准中取得开源第一，权重将在两周后开源。
GLM-5.3的公开信息集中于编程与网络安全表现。材料提到其在Terminal Bench 3.0等基准中取得开源第一，接近Claude Fable 5；白盒代码审查等安全任务表现持平Mythos 5，CyberGym得分为84.5%。
Aioga判断，此次发布的核心看点是后训练投入与代码、安全场景表现，而非基础模型更换。上述排名和接近关系依赖具体测试条件，仍需等待权重开放及独立复测验证。
若材料中的基准结果能够复现，GLM-5.3可能增强开源模型在编程代理和安全测试场景中的竞争力。权重开放后，开发者还需关注实际部署效果、工具接入体验与安全边界。 值得关注权重是否按材料所述在两周后开放，以及开源版本与ZCode、AutoClaw上线版本的能力和使用条件是否一致。后续应对照公开基准说明和独立测试，核验50%提升及84.5%得分。
The readable text on this page was extracted from the public source and organized with attribution, publication time and the original link. Copyright remains with the original author and publisher.
Ingestion channel: Summary aggregation · Source domain: mp.weixin.qq.com
Source: 公众号：智谱（GLM）
Original link: Open original source
Aioga archive: Open intelligence page
Content record: summary-fallback · Updated: 2026-08-14T05:31:04.000Z
统一接入主流 AI 模型 API，为开发、测试与生产环境提供稳定调用入口。
立即访问 api.w173.com
Scan to open this articleAioga aggregates global AI updates and preserves source information for verification and citation.