GLM-5.3 is now open-weight

Practical AI: Tools, Models & Frameworksopen-weight

What changed

GLM-5.3 uses the same base model as GLM-5.2 — every gain comes from post-training. It also achieve open-source SOTA on public benchmarks including Terminal Bench 3.0 and Agents' Last Exam. - NL2Repo: We evaluated NL2Repo with temperature=1.0 ,top_p=1.0 , andmax_new_tokens=64k under 1M context.

Why it matters

A concrete addition to Practical AI: Tools, Models & Frameworks: it changes what's available to builders today rather than being general commentary.

How it compares

Related prior coverage to compare against:

  • DeepSeek V4 Pro 0813
  • Nvidia, Microsoft, Meta warn against overregulating open-weight models
  • Running a 28.9M parameter LLM on an $8 microcontroller

Sources