GLM (AI)

GLM
DeveloperZ.ai
Release3 March 2021; 5 years ago (2021-03-03)
Stable release
GLM-5.3-FlashX
21 September 2026; 51 days ago (2026-09-21)
PlatformCloud computing platforms
Local computers
Type
LicenseVarious; see § Models

GLM (lit. 'General Language Model') is a series of open weight large language models developed by Chinese software company Z.ai.

History

GLM was first described in an 18 March 2021 paper titled GLM: General Language Model Pretraining with Autoregressive Blank Infilling.

GLM was released as a chatbot in March 2023 under the name ChatGLM.

GLM-5

Beginning with GLM-5, Z.ai adopted the DeepSeek Sparse Attention mechanism. The company's benchmarks show that the 5.2 iteration of the model released in June 2026 was nearly as performant as Claude Opus 4.7 to 4.8.

Z.ai released GLM-5.3 on 14 August 2026 and made the model's weights available two weeks later, it uses 753 billion parameters, with 40 billion active. In September, CAISI found GLM 5.3 to be the most advanced in cybersecurity among open-weight models. The same month, Anthropic's Frontier Red Team evaluated GLM-5.3's cybersecurity capability at a similar level to its Claude Mythos Preview, which it released to partner organizations in April 2026.

Z.ai released GLM-5.3-Flash on 26 August. As well as DeepSeek Sparse Attention, this model drew on Kimi Delta Attention from Chinese AI firm Moonshot AI. Z.ai disclosed that they had been serving the model using AI chips manufactured in China.

Models

List of models
Name Release date License Ref.
GLM March 2021 MIT
GLM-130B September 2022 Apache 2.0
ChatGLM-130B March 2023 Proprietary
ChatGLM-6B March 2023 Apache 2.0
ChatGLM2-6B June 2023
ChatGLM3-6B October 2023
GLM-4 (0116) January 2024 Proprietary
GLM-4 (0520) May 2024
GLM-4-Air (0605) June 2024
GLM-4-9B June 2024 Apache 2.0
GLM-4-Plus August 2024 Proprietary
GLM-4-Voice October 2024 Apache 2.0
GLM-4-Air-250414 April 2025 Proprietary
GLM-4-FlashX-250414 April 2025
GLM-4-Flash-250414 April 2025
GLM-4-9B-0414 April 2025 MIT
GLM-4-32B-0414 April 2025
GLM-4.1V-Thinking July 2025
GLM-4.5 July 2025
GLM-4.5V August 2025
GLM-4.6 September 2025
GLM-4.6V December 2025
GLM-4.7 December 2025
GLM-5 February 2026
GLM-5.1 April 2026
GLM-5.2 June 2026
GLM-5.3 August 2026 GLM-5.3
GLM-5.3-Flash August 2026 MIT
GLM-5.3-FlashX September 2026 Proprietary

Products

GLM is the flagship model series of Z.ai, one of the six AI tigers of China. The weights of most GLM models are released under the MIT License or Apache License 2.0, allowing them to be run either locally or in the cloud.

Though the first GLM model was published on 3 March 2021, it was released under the name ChatGLM (Chinese: 智譜清言) as an AI-based chatbot in March 2023. It is also used in AI-assisted software development.

In addition to the ability to run locally, such as via llama.cpp, GLM models are available via API at per-token costs directly from Z.ai. Z.ai also offers a subscription-based plan for API model access, the GLM Coding Plan. The plan is priced at approximately one tenth of a subscription to Claude Pro.

AutoGLM

Z.ai collaborated with Alibaba Cloud to release AutoGLM, a tool to run GLM from a mobile phone app similarly to OpenClaw.

Deployment

Hugging Face self-hosted GLM-5.2 for data processing to mitigate a cyberattack autonomously conducted by two OpenAI models, including GPT-5.6 Sol, that had escaped their sandbox in July 2026. Prior to deploying GLM-5.2, Hugging Face had attempted to use American proprietary frontier models, which were prevented by their AI safety guardrails from answering the company's requests.

Revolut switched its primary coding model from Claude to self-hosted GLM 5.1 (later GLM 5.3) to reduce coding time by 60% and token costs by 90%. The low costs made it affordable enough for coding models to be used frequently for debugging and solving bug tickets, instead of conserving tokens for the most difficult tasks. David Abitbol, the head of infrastructure and operations, acknowledged potential privacy and security issues of Chinese models but said that self-hosted models do not pose such problems.

See also

References

  1. ↑ Du, Zhengxiao; Qian, Yujie; Liu, Xiao; Ding, Ming; Qiu, Jiezhong; Yang, Zhilin; Tang, Jie (2021). "GLM: General Language Model Pretraining with Autoregressive Blank Infilling". arXiv:2103.10360 [cs.CL].
  2. ↑ ChatGLM：预训练大模型探索与实践 (PDF). tsinghua.edu (in Chinese).
  3. ↑ Zeng, Aohan; Lv, Xin; Hou, Zhenyu; Du, Zhengxiao; Zheng, Qinkai; Chen, Bin; Yin, Da; Ge, Chendi; et al. (24 February 2026). "GLM-5: from Vibe Coding to Agentic Engineering". arXiv:2602.15763 [cs.LG].
  4. ↑ "CEO of Chinese Anthropic rival tells Elon Musk that China will have a Fable 5-class AI model before next year — it 'won't take that long' says Jie Tang in response to Musk's prediction of a Q1 target". tomshardware.com. 19 June 2026.
  5. ↑ "GLM-5.3 (max) - Intelligence, Performance & Price Analysis | Artificial Analysis". artificialanalysis.ai. Retrieved 15 October 2026.
  6. 1 2 Baptista, Eduardo; Chen, Laurie (14 August 2026). "China's Z.ai says new model nears Anthropic's Mythos 5 in cyber-defence tests". Reuters. Retrieved 15 August 2026.
  7. 1 2 Kunz, Christopher (14 August 2026). "New version of the open-weight model: GLM 5.3 improves security capabilities". Heise Online. Retrieved 15 August 2026.
  8. 1 2 Lohr, Steve; Tobin, Meaghan (28 August 2026). "China Sees Opportunity in America's Recent A.I. Security Scares". The New York Times. Retrieved 28 August 2026.
  9. ↑ "CAISI's Assessment of Z.ai's GLM-5.3 Cyber Capabilities". NIST. 17 September 2026.
  10. ↑ Lee, Chong Ming (30 September 2026). "Anthropic raises alarm over Chinese GLM-5.3 model's elite hacking ability". South China Morning Post. Retrieved 30 October 2026.
  11. 1 2 Ding, Luz (26 August 2026). "China's Z.ai Made Ox Alpha Stealth Model, Rivaling DeepSeek". Bloomberg News. Retrieved 26 August 2026.
  12. ↑ "THUDM/GLM: GLM (General Language Model)". GitHub. Retrieved 17 July 2026.
  13. 1 2 3 "About us". zhipuai.cn.
  14. ↑ "zai-org/GLM-130B: GLM-130B: An Open Bilingual Pre-Trained Model (ICLR 2023)". GitHub. 17 July 2026. Retrieved 17 July 2026.
  15. 1 2 3 4 Wang; Zihan (18 June 2024). "ChatGLM: A Family of Large Language Models from GLM-130B to GLM-4 All Tools". arXiv:2406.12793 [cs.CL].
  16. ↑ "zai-org/chatglm-6b · Hugging Face". huggingface.co. 18 June 2024. Retrieved 17 July 2026.
  17. ↑ "zai-org/chatglm2-6b · Hugging Face". huggingface.co. 18 June 2024. Retrieved 17 July 2026.
  18. ↑ "zai-org/chatglm3-6b · Hugging Face". huggingface.co. 18 June 2024. Retrieved 17 July 2026.
  19. ↑ "zai-org/GLM-4: GLM-4 series: Open Multilingual Multimodal Chat LMs | 开源多语言多模态对话模型". GitHub. Retrieved 17 July 2026.
  20. 1 2 Nguyen, Britney (10 March 2025). "Meet the 'Six Tigers' that dominate China's AI industry". Quartz. Retrieved 12 April 2026.
  21. ↑ zai-org/GLM-4-Voice: GLM-4-Voice | 端到端中英语音对话模型. GitHub. Retrieved 17 July 2026.
  22. ↑ "LICENSE · zai-org/glm-4-voice-9b". Hugging Face. 25 October 2024. Retrieved 12 April 2026.
  23. 1 2 3 北京智谱华章科技有限公司 (3 October 2024). "ZHIPU AI OPEN PLATFORM". bigmodel.cn. Archived from the original on 3 October 2024. Retrieved 3 October 2024.
  24. ↑ "zai-org/GLM-4-9B-0414 · Hugging Face". huggingface.co. 2 March 2026. Retrieved 17 July 2026.
  25. ↑ "zai-org/GLM-4-32B-0414 · Hugging Face". huggingface.co. 2 March 2026. Retrieved 17 July 2026.
  26. ↑ Zhou, Wen (2 July 2025). 智谱获浦东创投、张江集团 10 亿元战略投资，开源发布新一代通用视觉语言模型 GLM-4.1V-Thinking [Zhipu Secures ¥1 Billion in Strategic Investment from Pudong Venture Capital and Zhangjiang Group, Open-Sources Next-Generation General-Purpose Vision-Language Model GLM-4.1V-Thinking]. ITHome (in Chinese). Retrieved 12 April 2026.
  27. ↑ "README.md · zai-org/GLM-4.1V-9B-Thinking". 25 October 2025. Retrieved 12 April 2026 – via Hugging Face.
  28. ↑ Cheng, Evelyn (28 July 2025). "China's latest AI model claims to be even cheaper to use than DeepSeek". CNBC. Retrieved 19 August 2025.
  29. ↑ "README.md · zai-org/GLM-4.5". 11 August 2025. Retrieved 12 April 2026 – via Hugging Face.
  30. ↑ Hao Miao (11 August 2025). 智谱视觉推理模型 GLM-4.5V 上线并开源，号称“全球 100B 级效果最佳” [Zhipu's Vision-Language Model GLM-4.5V Launches and Is Open-Sourced, Hailed as the "Best-Performing 100B-Class Model Globally"]. ITHome (in Chinese). Retrieved 11 April 2026.
  31. ↑ "README.md · zai-org/GLM-4.5V". 25 October 2025. Retrieved 11 April 2026 – via Hugging Face.
  32. ↑ Chow, Vincent (1 October 2025). "China's Z.ai targets US AI leaders with new flagship model as coding wars heat up". South China Morning Post. Retrieved 11 April 2026.
  33. ↑ "README.md · zai-org/GLM-4.6". 30 September 2025. Retrieved 11 April 2026 – via Hugging Face.
  34. ↑ Franzen, Carl (8 December 2025). "Z.ai debuts open source GLM-4.6V, a native tool-calling vision model for multimodal reasoning". VentureBeat. Retrieved 11 April 2026.
  35. ↑ "README.md · zai-org/GLM-4.6V". 9 December 2025. Retrieved 11 April 2026 – via Hugging Face.
  36. ↑ "Introducing the Chinese-made AI 'GLM-4.7,' which is strong in coding, and an open model that surpasses Gemini 3.0 Pro in some tests". Gigazine. 23 December 2025. Retrieved 11 April 2026.
  37. ↑ "README.md · zai-org/GLM-4.7". 8 January 2026. Retrieved 11 April 2026 – via Hugging Face.
  38. ↑ Chow, Vincent (12 February 2026). "China's Zhipu AI launches new major model GLM-5 in challenge to its rivals". South China Morning Post. Retrieved 17 February 2026.
  39. ↑ Mo, Liam; Goh, Brenda; Chen, Laurie (11 February 2026). D'Silva, Anil (ed.). "Chinese AI startup Zhipu releases new flagship model GLM-5". Reuters. Retrieved 17 February 2026.
  40. ↑ "README.md · zai-org/GLM-5". 5 April 2026. Retrieved 11 April 2026 – via Hugging Face.
  41. ↑ Chang, Minxiao (8 April 2026). "China's Zhipu AI open-sources flagship model, raises prices to narrow gap with US rivals". South China Morning Post. Archived from the original on 9 April 2026. Retrieved 10 April 2026.
  42. ↑ Thomas, Prasanth Aby (8 April 2026). "Z.ai unveils GLM-5.1, enabling AI coding agents to run autonomously for hours". Computerworld. Retrieved 11 April 2026.
  43. ↑ "README.md · zai-org/GLM-5.1". 8 April 2026. Retrieved 11 April 2026 – via Hugging Face.
  44. 1 2 Chang, Minxiao (15 June 2026). "Zhipu AI's stock rockets after Chinese firm launches GLM-5.2 model". South China Morning Post. Retrieved 17 June 2026.
  45. ↑ "China's Z.ai GLM-5.2 tops OpenAI's GPT 5.5 model on key benchmarks". The Economic Times. 17 June 2026. Retrieved 17 June 2026.
  46. ↑ "GLM-5.2: Built for Long-Horizon Tasks". Z.ai. Retrieved 16 June 2026.
  47. ↑ "LICENSE · zai-org/GLM-5.3". 28 August 2026. Retrieved 28 August 2026 – via Hugging Face.
  48. ↑ "LICENSE · zai-org/GLM-5.3-Flash". 26 August 2026. Retrieved 26 August 2026 – via Hugging Face.
  49. ↑ "Z.ai API Platform — Start building with GLM-5.3". z.ai. Retrieved 20 August 2026.
  50. ↑ "OpenClaw sparks boom as Chinese firms race into the AI agent era". technode.com. 10 March 2026.
  51. ↑ Soni, Aditya; Singh, Jaspreet (22 July 2026). "Chinese AI's role in stopping rogue OpenAI agent shows cost of US guardrails". Reuters. Retrieved 23 July 2026.
  52. ↑ Goudarzi, Sara (23 July 2026). "What the OpenAI-Hugging Face incident reveals about US-China AI interdependence". Bulletin of the Atomic Scientists. Retrieved 23 July 2026.
  53. ↑ Shen, Xinmei (22 July 2026). "China's Zhipu AI model contains hack after OpenAI models go rogue". South China Morning Post. Retrieved 23 July 2026.
  54. ↑ Mann, Tobias (3 November 2026). "British fintech Revolut slashes AI coding costs 10x by cutting out Claude and going self-hosted". theregister. Retrieved 10 November 2026.
  55. ↑ Finance, BigGo (29 October 2026). "Revolut's David Abitbol Says Self-Hosted GLM Models Cut AI Coding Costs 90% for 2,000 Engineers". BigGo Finance. Retrieved 10 November 2026.