avatar

Z.ai Opens GLM-5.1 to All Coding Plan Subscribers

Chinese AI company makes its 744B-parameter model available at $10/month, undercutting Western competitors

avatar@Zai_org
5 months ago

TL;DR:

  • Z.ai's GLM-5.1, a 744-billion parameter model, now available to all GLM Coding Plan tiers starting around $10/month
  • The model scores 77.8% on SWE-bench Verified, close to Claude Opus and ahead of Gemini 3 Pro
  • Subscription offers 3-5x more usage than comparable Anthropic plans, with support for Claude Code and Cursor
  • Z.ai runs on Huawei Ascend chips and releases under MIT license, positioning itself against U.S. closed-source models

Headline

Z.ai Opens GLM-5.1 Model to All Coding Plan Subscribers

Summary

Z.ai, a Tsinghua University spinout founded in 2019 and listed on the Hong Kong Stock Exchange in January 2026, announced that GLM-5.1 is now available across all tiers of its coding subscription service. The model has 744 billion parameters using a Mixture-of-Experts architecture.

Pricing starts around $10/month (Lite at $27/quarter, Pro at $81/quarter, Max at $216/quarter). Z.ai claims the plans offer 3-5x the usage of Anthropic's Claude Pro. The service works with Claude Code and Cursor.

GLM-5.1 scores 77.8% on SWE-bench Verified, beating Gemini 3 Pro and approaching Claude Opus 4.5. The model is open-source under MIT license.

Analysis

This pricing move puts pressure on Western AI companies. Z.ai is betting that developers will switch for the cost savings, especially those running long coding sessions where token costs add up.

The company runs on Huawei Ascend chips rather than Nvidia hardware, which matters given ongoing U.S. export restrictions. That's a selling point for customers worried about supply chain issues.

Developers on forums have noted lower hallucination rates compared to GPT-4.5, though independent benchmarks are still limited. The flat subscription model (rather than per-token pricing) removes unpredictable costs that frustrate many developers.

For Anthropic and OpenAI, this creates an awkward choice: match the aggressive pricing or differentiate on quality and reliability. Early signs suggest they're betting on the latter.

Impact Assessment

  • Significance: High
  • Categories: Model Release, Developer Tools, Market Impact