Z.ai Launched GLM-4.6, Delivering Breakthroughs in Coding Capabilities

Z.ai recently announced the official release of GLM-4.6, the latest model in its GLM series. This new model delivers comprehensive enhancements across real-world coding, long-context processing, reasoning, and agentic applications, marking another significant milestone for Chinese models.

GLM-4.6 is now live on the Z.ai Chat [http://chat.z.ai] and the Z.ai API platform [https://docs.z.ai/guides/llm/glm-4.6], and it has also been open-sourced on Hugging Face [https://huggingface.co/zai-org/GLM-4.6] under the MIT license.

Comprehensive Upgrade of Core Capabilities

GLM-4.6 showcases significant improvements across several key dimensions. The context window has been expanded from 128K to 200K tokens, enabling it to handle more complex agentic tasks. In its core coding capabilities, the model not only achieves higher scores on standard benchmarks but also excels in real-world applications such as Claude Code, Cline, and Roo Code.

Furthermore, the reasoning performance of GLM-4.6 has been significantly improved and now supports tool use. Its agentic and writing capabilities have also been refined, allowing for more effective framework integration and text generation that better aligns with human preferences.

Performance Evaluation: A New High for Open-Weight Models

In performance evaluations, GLM-4.6 achieves results on par with Claude Sonnet 4 and Claude Sonnet 4.5 across 8 authoritative general capability benchmarks, including AIME 25, GPQA, and SWE-Bench Verified.

Image: https://www.globalnewslines.com/uploads/2025/10/c783e78722501d3b8b933274970df128.jpg

To more accurately assess its practical programming skills, Z.ai conducted 74 real-world coding tests within the Claude Code environment, where the results showed that GLM-4.6 surpassed the performance of Claude Sonnet 4 and other open-weight models.

Image: https://www.globalnewslines.com/uploads/2025/10/1d0f4c463ee5dcf360b303e7838a5c28.jpg

In terms of efficiency, the model's average token consumption has been reduced by over 30% compared to its predecessor, GLM-4.5. To ensure transparency and credibility, all test data has been made publicly available on Hugging Face for external verification and reproduction.

Image: https://www.globalnewslines.com/uploads/2025/10/e36351ca6feaca3f0fbd4bbd97fe03dd.jpg

Major Upgrade to the GLM Coding Plan

Alongside the new model, Z.ai is simultaneously implementing a major upgrade to its GLM Coding Plan [https://z.ai/subscribe], which aims to provide users with a top-tier coding experience at a highly competitive price. Existing subscribers will be automatically upgraded to GLM-4.6 and will gain access to new features, including image recognition, search capabilities, and support for over 10 mainstream coding agents. For developers with high-volume needs, Z.ai is also introducing the GLM Coding Max plan, offering three times the usage of the Claude Max (20x) plan.

Continuing the Growth Momentum

The release of GLM-4.6 builds upon the success of its predecessor, GLM-4.5, which was the first single model to natively integrate reasoning, coding, and agentic capabilities. Following its release, commercial API usage on the Z.ai API platform grew by more than tenfold. For more technical details, please refer to the GLM-4.6 technical blog [https://z.ai/blog/glm-4.6].
Media Contact
Company Name: Z.ai
Contact Person: Zixuan Li
Email: Send Email [http://www.universalpressrelease.com/?pr=zai-launched-glm46-delivering-breakthroughs-in-coding-capabilities]
Country: Singapore
Website: https://z.ai/subscribe

Legal Disclaimer: Information contained on this page is provided by an independent third-party content provider. GetNews makes no warranties or responsibility or liability for the accuracy, content, images, videos, licenses, completeness, legality, or reliability of the information contained in this article. If you are affiliated with this article or have any complaints or copyright issues related to this article and would like it to be removed, please contact retract@swscontact.com



This release was published on openPR.