CPU-only GLM 5.2: Epyc and 512GB RAM
This is just a preview of some content I'm putting together to share with you all. I have a server I've put together and I'm testing…
GLM 5.2 Q1_S vs Qwen 27B Q8
TL;DR; GLM-5.2 Q1_S beats Qwen 3.6 27B Q8, both run at KV Q8 Disclaimer: This is a hobby/amateur comparison with n=1, so go easy on it.…
RT jietang: Any new features we must have in the next version of glm?
RT jietangAny new features we must have in the next version of glm?
High-quality GLM-5.2 Quant on 4x DGX Spark – Guide, Results, and Comps
I got GLM-5.2 NVFP4 running on four DGX Sparks at 128K context. This is still a niche/hacky setup, but it is now a real serving point…
China’s Z.ai claims it can match Mythos on cybersecurity
China's Zhipu AI (Z.ai) released its open-weight GLM-5.2, and some researchers have claimed that it matches Mythos in certain bug-finding and cybersecurity scenarios. While GLM lags…
More reason why we’re excited about GLM-5.2 on Together 👇 Strong enough for serious coding work, cheap enough to change routing decisions, and eas…
More reason why we’re excited about GLM-5.2 on Together 👇Strong enough for serious coding work, cheap enough to change routing decisions, and easy to access through…
Semgrep: GLM 5.2 beats Claude in our Cyber Benchmarks
Article URL: https://semgrep.dev/blog/2026/we-have-mythos-at-home-glm-52-beats-claude-in-our-cyber-benchmarks/ Comments URL: https://news.ycombinator.com/item?id=48709670 Points: 6 # Comments: 1
RT Yuchen Jin: GLM-5.2 is the open-source Claude moment. The demand we’re seeing at Databricks is astonishing. The world is going to see massive adop…
RT Yuchen JinGLM-5.2 is the open-source Claude moment.The demand we’re seeing at Databricks is astonishing. The world is going to see massive adoption of oss LLMs.Also,…
RT Janek Mann: There's a bogus headline making the rounds claiming "GLM-5.2 beats Mythos in security bug detection", but it's not vs Mythos but vs Opu…
RT Janek MannThere's a bogus headline making the rounds claiming "GLM-5.2 beats Mythos in security bug detection", but it's not vs Mythos but vs Opus, not…
What is this slop? Do they actually mean GLM 5.2-Cyber (non-nerfed version), or some unrepresentative eval?
What is this slop? Do they actually mean GLM 5.2-Cyber (non-nerfed version), or some unrepresentative eval?First Squawk: ZHIPU AI’S NEW MODEL REPORTEDLY MATCHES CLAUDE MYTHOS IN…
RT xjdr: we've been running GLM 5.2 in bf16 and in fp8 (experts and kvcache only, attention is always bf16) and have recorded virtually 0 measurable q…
RT xjdrwe've been running GLM 5.2 in bf16 and in fp8 (experts and kvcache only, attention is always bf16) and have recorded virtually 0 measurable quality…
RT Alexander Doria: I mean just HF out, ModelScope blocked and it’s over. You won’t seed GLM.
RT Alexander DoriaI mean just HF out, ModelScope blocked and it’s over. You won’t seed GLM.Naithan Jones: “THeY cAnT bAn OpEn sOuRcE bRO hOw cAn tHeY…
Running GLM5.2 on budget hardware < $2500.
Too many times I hear people whine about not being ble to run SOTA models or claim it would require $50k, or $100k. https://www.ebay.com/itm/398079051468 Epcy Motherboard…
Honestly this makes the whole benchmark look even more absurd. Grok 4.20 over Opus 4.8 (max), Kimi K2.5 > GLM 5.2 and Opus 4.7, Opus 4.6 down in the d…
Honestly this makes the whole benchmark look even more absurd. Grok 4.20 over Opus 4.8 (max), Kimi K2.5 > GLM 5.2 and Opus 4.7, Opus 4.6…
I actually think that if recent rumors about GLM being propped up by distillation (or rather, Claude Code use interception) are correct, this is very …
I actually think that if recent rumors about GLM being propped up by distillation (or rather, Claude Code use interception) are correct, this is very *bullish*…
RT Maziyar PANAHI: Got GLM-5.2 running on my Mac Studio via llama.cpp, the reasoning behind all my medical agentic workflows. It orchestrates a swarm …
RT Maziyar PANAHIGot GLM-5.2 running on my Mac Studio via llama.cpp, the reasoning behind all my medical agentic workflows.It orchestrates a swarm of tiny on-device OpenMed…
> What makes GLM 5.2 interesting: zero failed runs across 84 runs (vs ~10% failure rate for Opus agents). The most reliable agent we've seen The wonde…
> What makes GLM 5.2 interesting: zero failed runs across 84 runs (vs ~10% failure rate for Opus agents). The most reliable agent we've seenThe wonders…
As token usage explodes, model choice becomes product strategy. Teams are already testing models like GLM-5.2 because they want frontier quality, bett…
As token usage explodes, model choice becomes product strategy.Teams are already testing models like GLM-5.2 because they want frontier quality, better tokenomics, and more control over…
RT DailyPapers: NVIDIA just released an optimized GLM-5.2 on Hugging Face A 753B parameter MoE with 1M context, quantized to NVFP4 for Blackwell GPUs…
RT DailyPapersNVIDIA just released an optimized GLM-5.2 on Hugging FaceA 753B parameter MoE with 1M context,quantized to NVFP4 for Blackwell GPUs—nearly matching FP8 accuracy.
Qwen 3.6 27b GLM 5.2 fine-tune?
Hi everyone, Since both models are open weights and GLM seems to find that secret to frontier model reasoning, why don't we see any Qwen GLM…
RT Hassan: I love using GLM 5.2 for web app iteration. My workflow: generate 6 variations, then pick the best one and continue iterating on it. I buil…
RT HassanI love using GLM 5.2 for web app iteration.My workflow: generate 6 variations, then pick the best one and continue iterating on it.I built Recast…
RT Victor M: Cool: if you have a Hugging Face account you have enough free credits to ask GLM-5.2 to build your website on HuggingChat (it will use Ex…
RT Victor MCool: if you have a Hugging Face account you have enough free credits to ask GLM-5.2 to build your website on HuggingChat (it will…
Does this answer your question about its positioning @TheZvi ? I told you it's a hardware issue. GLM 5.2 can be served very, very quickly. Databricks …
Does this answer your question about its positioning @TheZvi ? I told you it's a hardware issue. GLM 5.2 can be served very, very quickly. Databricks…
RT Niels Rogge: Impressive release by @ornith_! Getting the top spot on Terminal Bench 2.1 with only GLM-5.2 (a much bigger model) above it! Also beat…
RT Niels RoggeImpressive release by @ornith_!Getting the top spot on Terminal Bench 2.1 with only GLM-5.2 (a much bigger model) above it!Also beating Opus 4.8 👀Ornith:…