# GLM-5

> **TL;DR.** GLM-5 is Zhipu AI's (Z.

- **Category:** AI / LLMs / Open Source
- **Stage:** rising
- **Age:** 149 days
- **Origin date:** 2026-02-11
- **First detected:** 2026-04-30
- **Canonical URL:** https://earlyterms.com/term/glm5
- **Sources:** 9 primary URLs

## Definition

GLM-5 is Zhipu AI's (Z.ai) flagship open-weight large language model, a 744-billion-parameter Mixture-of-Experts system built for agentic coding and long-horizon tool use. Released under the MIT license, it anchors a fast-iterating family that spawned GLM-5.1 and GLM-5.2 within four months.

It launched on [February 11, 2026](https://z.ai/blog/glm-5), trained entirely on 100,000 Huawei Ascend chips using Huawei's MindSpore framework, with zero Nvidia dependency — a first for a frontier-class model. The release landed weeks after Zhipu's Hong Kong IPO; shares jumped 34% that day, and SWE-bench Verified hit 77.8%.

## Analogy

GLM-5 is China's proof that a frontier model can be forged on homegrown chips instead of Nvidia silicon.

## Why it's emerging now

Zhipu AI's GLM-5 family became the reference open-weight model for agentic coding after its Feb 11, 2026 launch on domestic Huawei Ascend chips, and Zhipu's Hong Kong-listed stock has ridden each GLM-5.x release to a market cap over HK$1 trillion.

## Related terms

- *child:* GLM-5.1
- *child:* GLM-5.2
- *related:* GLM-4.7
- *competitor:* DeepSeek V4 Pro
- *competitor:* Qwen3.6
- *competitor:* Kimi K2.6
- *competitor:* Claude Opus 4.8
- *competitor:* GPT-5.5
- *competitor:* Grok 4.5
- *competitor:* Mythos 5
- *related:* coding plan
- *alias:* 智谱GLM-5

## Sources

1. [Z.ai — GLM-5: From Vibe Coding to Agentic Engineering](https://z.ai/blog/glm-5)
2. [zai-org/GLM-5 (GitHub repo)](https://github.com/zai-org/GLM-5)
3. [zai-org/GLM-5 (Hugging Face model card)](https://huggingface.co/zai-org/GLM-5)
4. [Hugging Face blog (Maxime Labonne) — GLM-5: China's First Public AI Company Ships a Frontier Model](https://huggingface.co/blog/mlabonne/glm-5)
5. [Let's Data Science — How China's GLM-5 Works: 744B Model on Huawei Chips](https://letsdatascience.com/blog/china-trained-frontier-ai-model-glm-5-without-nvidia)
6. [South China Morning Post — Zhipu AI market cap tops HK$1 trillion as GLM-5.2 developer soars](https://www.scmp.com/tech/article/3357858/zhipu-ai-market-cap-tops-hk1-trillion-shares-glm-52-developer-soar)
7. [Semgrep — GLM 5.2 beats Claude in our benchmarks](https://semgrep.dev/blog/2026/we-have-mythos-at-home-glm-52-beats-claude-in-our-cyber-benchmarks/)
8. [Hacker News — GLM 5.2 beats Claude in our benchmarks (1,113 points)](https://news.ycombinator.com/item?id=48709670)
9. [Artificial Analysis — GLM-5.2 is the new leading open weights model on the Intelligence Index](https://artificialanalysis.ai/articles/glm-5-2-is-the-new-leading-open-weights-model-on-the-artificial-analysis-intelligence-index)

---
_Generated by EarlyTerms · https://earlyterms.com/term/glm5_