Event ·Multi-source

智谱发布GLM-5.3，自称最强开源权重编码模型

智谱AI发布GLM-5.3，并称其为最强的开源权重编码模型。据其自家基准，仅通过后训练就比上一代提升50%。该模型经过网络安全训练，帮助安全团队在269个项目中找到2436个漏洞。模型权重计划在两周后开源。

Current read

智谱发了GLM-5.3，说是开源编码模型里最强的，后训练提升50%，还帮找出了2436个漏洞，权重两周后开源，可以关注下。

11 sources72° AI heatLast updated 2026/8/14 10:21:34

Evidence

11 sources

Primary sourceDecoder

2026/8/14 10:21:34View source

Related source 1IT之家

2026/8/14 05:36:12View source

Related source 2向阳乔木

2026/8/14 08:15:58View source

Related source 3Aravind Srinivas

2026/8/14 15:01:20View source

Related source 4elvis

2026/8/14 15:23:08View source

Related source 5歸藏(guizang.ai)

2026/8/14 16:39:05View source

Related source 6lmarena.ai

2026/8/14 17:59:52View source

Related source 7宝玉

2026/8/14 05:50:41View source

Related source 8Geek

2026/8/14 05:55:55View source

Related source 9ollama

2026/8/14 06:37:30View source

Related source 10Pandaily

2026/8/14 08:12:12View source

Conflict check

Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.

Open the story