Vibe CodingGLM-5 744B Open Source Free on NVIDIA NIM 77.8 Percent SWE-benchNVIDIA NIM·high signalXBlueskyLinkedInCopy linkGLM-5 744B/40B MoE, MIT license, free on NVIDIA NIM. 77.8% SWE-bench (highest open-source). Trained on Huawei Ascend.SourceSource pageNVIDIA NIM↳ Follow the threadShared entity / Actor roleParitok-4B compresses coding-agent context to 25.7% of size with a 264 MB adapter, and shows gpt-5 as a compressor is net-negativearXivStack layer / Threat patternCROSS-CATEGORY: The Guardrail Layer Around Coding Agents Shipped at Four Different Points in the Lifecycle in 48 HoursHacker News Show HN and Product Hunt (2026-08-24 and 2026-08-25)Stack layer / ContrastOrnith-1.5 Ships 9B Dense, 35B MoE and 397B MoE Under MIT, With the 397B Claiming 86.0 SWE-bench Verified Against Claude Opus 4.8's 85.8Hugging FaceStack layer / Contrastd-Matrix Raptor Claims 100 TB/s From 3D-DRAM and 20x the Bandwidth Density of HBMServeTheHomeStack layer / ContrastCline v4.1.16 fixes hooks resolving from global state, and starts redacting credentials embedded in git remote URLs before they reach the modelGitHub (cline/cline)Policy dependency / Threat patternmcp-shell Ships Security Off in One Deploy Path and Bypassable in the Other (CVE-2026-55580/55581/55582)GitHub Security AdvisoriesStack layer / ContrastVercel Puts MiniMax M3 and M2.7 on AI Gateway Free Through September 6Vercel ChangelogStack layer / Update threadIBM Releases Granite Speech 5.0 TurboCTC, a 470M Apache-2.0 ASR Model Claiming 12,600 RTFx on One H200IBM Granite (Hugging Face)