랭킹 > Z.ai > GLM-5.3
GLM-5.3
최근 평가: 대체로 긍정적 · 긍정 70% · 최근 평가 43개
개발사: Z.ai · 출시일: 2026-08-14
- 추천 · @akses_0x00: I think there’s a myth forming that you need an abliterated version of GLM5.3 to do vulnerability research or exploitability analysis/proof. That has not been my experience at all. Stock works fine for me? What are yall asking for that it refuses ?? (Don’t answer that 😅)
- 추천 · @Moham88Ab: هجوم برمجي وراءه إيران وتم استخدام موديل GLM 5.3 لتنفيذ الهجوم ، احنا في شركتنا بنشغل الموديل ده محلياً وعارفين إمكانياته كويس ويقدر يخترق البنوك وشركات الطيران ويفشخ حركة المرور
- 추천 · @_BILLDING_: here's the model breakdown. Pretty heavy on GLM 5.3 but I think it's my favorite model anyway. I really just have a nice setup in discord lol, one channel for every agent, a thread for crons for each one, and then just open up threads and close them as needed to keep only the th…
- 비추천 · @Ursa81370002: 我感觉我的包年智谱 Coding Plan Max 白买了，一个 bug GPT-6-Astra 2分钟就找到并完美复现了，GLM-5.3 花了几小时找不出来还给了个方向完全错误的方案。
- 비추천 · @Apuipuic: dsh再起動かけてもURL変わらんから開けるよ〜ってGLM5.3に言われたから信じて許可したら出先で開けなくなって詰んだ
- 비추천 · @cryptolukesky: Pi 加 DeepSeek-V4.1-Flash，让它起几个子代理，用这两个 5.3 也去分析一下。好家伙，一个小时过去我一看它花了我快 50 美元的 GLM-5.3 的 token 费。一问它，好家伙，它写了个 bug 出来，导致每个任务重复调用了好几次。 所以说。日常优先用SOTA会让你的减少很多不必要的麻烦，甚至是损失