第一项始终是图钉的来源。总体可信度是各份资料对上方所用开始和结束时间支持程度的加权平均;资料每比最新的一份旧 180 天,权重减半。
Z.ai 于 2026年8月14日 将 GLM-5.3 上线至其 API、GLM Coding Plan 和 ZCode 客户端:路透社(经 The Standard[6] 转载)的报道日期为8月14日,称 Z.ai“周五表示”,DataNorth[5] 则称其“于 2026年8月14日 发布”。Z.ai 的开发者发布说明将该条目记在8月18日,开放权重于8月28日随后发布;这两者都是后续步骤,而非发布本身。
Z.ai's pricing page lists GLM-5.3 at $1.4 per million input tokens, $0.26 cached input and $4.4 output, the same as GLM-5.2.[1]
Reports Anthropic's Frontier Red Team analysis of 30 September (50 of 410 ExploitBench attempts against 56 for Claude Mythos Preview; safeguards bypassed in 64-100% of simulated tries), the new licence for companies above $10 billion revenue, and NIST CAISI calling it the most cyber-capable open-weight model but about four months behind US frontier models.[1]
Z.ai's model card: same base model as GLM-5.2 with every gain from post-training, the full results table against GLM-5.2, Kimi K3, DeepSeek-V4 Pro, Qwen3.8-Max, Claude Opus 4.8, Fable 5 and GPT-5.6 Sol, and the custom glm-5.3 licence; the repository went public on 27-28 August 2026.[1]
DataNorth's 17 August write-up: released 14 August, 1M context and 128K output, the Terminal-Bench 3.0 harness details, which results Artificial Analysis and Proximal ran, and the Coding Plan at $18 per month with half-price off-peak points.
有遗漏或错误吗?用你自己的话说明:能佐证此图钉的链接、不同的开始或结束日期及理由,或缺失、有误的信息。AI 会对照此图钉的来源进行核实,搜索更好的来源,并添加任何支持你说法的页面。图钉自身的来源仍然最重要。AI 也会查看图片:显示的是别的东西或显示效果差的图片会被移到后面或替换。
原标题: GLM-5.3 Released
Z.ai模型卡中发布的结果表(原样转载)[4];非 Z.ai 的列为各模型厂商自己的对比测试结果。
| 基准 | GLM-5.3 | GLM-5.2 | Kimi K3 | DeepSeek-V4 Pro-0813 | Qwen3.8-Max | Opus 4.8 | Fable 5 (含回退) | GPT-5.6 Sol |
|---|---|---|---|---|---|---|---|---|
| Terminal Bench 2.1 | 88.2 | 81.0 | 88.3 | 87.9 | 86.6 | 85.0 | 88.0 | 88.8 |
| Terminal Bench 3.0 | 28.3 | 4.6 | 17.4 | – | – | 21.1 | 33.7 | 34.6 |
| DeepSWE (v1.1) | 66.9 | 46.2 | 67.5 | 62.7 | 56.6 | 58.0 | 69.7 | 72.7 |
| NL2Repo | 58.0 | 48.9 | 58.0 | 61.1 | 55.9 | 69.7 | – | – |
| ProgramBench (几乎解决) | 19.0 | 9.5 | 17.5 | – | 10.5 | 15.5 | 33.0 | 23.0 |
| FrontierSWE | 78.1 | 67.5 | – | – | – | 66.5 | 88.2 | – |
| SWE-Marathon (v1.1) | 42.5 | 19.4 | 48.1 | – | – | 48.8 | 33.1 | 42.5 |
| PostTrainBench | 39.8 | 31.7 | 32.0 | – | – | 32.9 | 41.8 | 36.2 |
| CyberGym | 84.5 | 77.2 | 80.0 | 83.3 | 78.5 | 78.1 | 83.8 | 83.6 |
| ExploitGym (2h / 6h) | 105 / 130 | 29 / 39 | 36 / 70 | – | 14 / 26 | 80 / 120 | 181 / 247 | 216 / 293 |
| ExploitBench | 54.4 | 24.4 | 32.2 | – | 28.8 | 40.0 | 78.0 | 76.5 |
| Toolathlon Verified | 73.0 | 59.9 | 76.5 | 74.1 | 72.5 | 76.2 | 74.7 | 74.9 |
| AutomationBench (v1.0.6) | 48.2 | 26.2 | 46.7 | 43.2 | 39.8 | 41.0 | 46.2 | 45.8 |
| Agents' Last Exam (ALE-CLI) | 28.5 | 23.8 | 27.6 | 25.7 | 27.0 | 25.7 | 23.8 | 28.6 |
| HLE w/ Tools | 62.5 | 54.7 | 59.8 | 60.0 | 56.2 | 57.9 | 63.9 | 64.5 |
| GDPval-AA v2 | 1769 | 1508 | 1682 | 1590 | 1739 | 1588 | 1743 | 1730 |
GDPval-AA v2 由 Artificial Analysis 运行,FrontierSWE 由 Proximal 运行,Toolathlon Verified 则通过该基准的官方服务运行[5].