model roundup

GLM 5.1

4 items · started 2026-04-25 · closed 2026-05-04

Built a self-hosted agent for small businesses that writes its own skills. ~$0.15 per customer booking on GLM-5.1 (www.reddit.com)

+14 7w glm

Been working on this for a while and finally at a point where it's running in production for a couple of small businesses, so figured I'd share. The thing that kept bugging me about "AI employee" products is that none of them are something…
Ask HN: Are there any good open-source chat apps? (news.ycombinator.com)

+2 8w glm ollama sonnet

Hi HN family! I've recently been messing around with open models through ollama (glm-5.1 and kimi-k2.6), and I've been impressed with just how close they are to Claude Sonnet for my needs, especially programming.
GLM 5.1 Locally: 40tps, 2000+ pp/s (www.reddit.com)

+78 8w glm sonnet claude-code

After some sglang patching and countless experiments, managed to get reap-ed nvfp4 version running stable and FAST on 4 x RTX 6000 Pros (limited to 350W). Very happy with performance and quality.
Capacity vs Speed trade-off: 1.1TB Mac Unified Memory vs. RTX 6000 Pros (www.reddit.com)

+38 9w glm

I'm usually a Windows person, but I’m currently running a Mac cluster for local LLM orchestration. My setup consists of four 256GB Mac Studios plus one 96GB Mac Studio, giving me about 1.1TB of unified memory.