Zibuu AI issues GLM-4.6: Finding enhancements in real codes in the world, processing a long context, consultation, search and Agentic AI

Zibuu Ai released GLM-4.6, a large update in its GLM series focus on the travel service in Agentic, to think of long-term content, and active coding activities. The model lifts the install window to 200k tokens with a 128k max outputintended for the use of low token in used work, and ships with Racked instruments of the shipment of the area.

So, what's new?

• Restricts + Output: 200 The input context and 123k Token tokens are high.
• The Real Secretary Consequences: Extended CC-Bench (Various conventions conducted by people's analysts in remote docket areas), reported GLM-4.6 closely with parttity with a claude summed 4 (48.6% winning a measure) and uses ~ 15% tokens a few tokens vs gl. GLM-4.5 to complete tasks. Task Prompts and Agent Trajects published for inspection.
• Benchmark position: Zubu is shortening “to achieve clear” on top of GLM-4.5 at all eight Benches of the public and the Stity Party and Claude Sonnet 4/16 for many; Reinstalls GLM-4.6 remains for Lags Sonnet 4.5 in codesA useful caveat of model is selective.
• Ecosystem availability: GLM-4.6 is available with Z.ai api including Opencrouter; It meets popular codes (Claude Code, Caline, ROO code, Kilo code), and existing strategy users can update the model name at the glm-4.6.
• Open license + license: To enter a list of facial model cards License: Mitha including Model size: 355B Param (MOE) With BF16 / F32 Tesons. .
• Diagnosis: vllm including Shang Supported by local service; The instruments are open Kisses face including Models.

Summary

GLM-4.6 Divorceful But Visible Step:

Kilombo

1) What is the Topic Limit?
GLM-4.6 supports a 200 The input context and 123k Token tokens are high.

2) Open instruments found and in what license?
Yes. List of Hugging Faculty Cards Racked instruments reference License: Mitha as well as a 357B-parameter Moe Configuration (BF16 / F32 tessors).

3) How does GLM-4.6 compare with GLM-4.5 and Cloud Sonnet 4 in used activities?
Extended CC-BenchGLM-4.6 reports ~ 15% tokens a few tokens vs gl. GLM-4.5 including near parity with Claude Sonnet 4 (48.6% Win-Rate).

4) Can I use GLM-4.6 in the area?
Yes. Zubuu provides instruments Tying face / models and local acquisition documents with vllm including Shang; Community values ​​appear in Workstation-Class Hardware.

Look GitHub page, Hugging Face Model Card including Technical Details. Feel free to look our GITHUB page for tutorials, codes and letters of writing. Also, feel free to follow it Sane and don't forget to join ours 100K + ml subreddit Then sign up for Our newspaper.

Asphazzaq is a Markteach Media Inc. According to a View Business and Developer, Asifi is committed to integrating a good social intelligence. His latest attempt is launched by the launch of the chemistrylife plan for an intelligence, MarktechPost, a devastating intimate practice of a machine learning and deep learning issues that are clearly and easily understood. The platform is adhering to more than two million moon visits, indicating its popularity between the audience.