This article first appeared on GuruFocus .
Zhipu is beginning to lean into monetization after a prolonged investment phase, raising prices for access to its most advanced AI models by at least 8% alongside the release of its latest open-source system, GLM-5.1. The Beijing-based company has adjusted cloud usage pricing, including input tokens, to levels roughly 8% to 17% higher than its prior GLM-5 Turbo model, marking another increase in 2026 following earlier adjustments. The move comes as demand for more advanced agentic AI services continues to build, while investors appear increasingly focused on whether leading developers can convert scale and capability into more durable revenue streams.
The pricing shift places Zhipu alongside domestic platforms such as Alibaba Group Holding Ltd. ( NYSE:BABA ) and Tencent Holdings Ltd. ( TCEHY ), which have also been lifting charges amid rising adoption of AI-driven services. The newly released GLM-5.1 introduces upgrades in coding and long-horizon task execution, with the ability to run autonomously for up to eight hours, and the company indicated its coding performance approaches Anthropic's Claude Opus 4.6. Chief Executive Officer Zhang Peng said he expects exponential growth in the firm's cloud-based models business, suggesting pricing could become an increasingly important lever as usage expands.
That backdrop sits against still-elevated losses, with Zhipu reporting a wider deficit of 4.7 billion yuan for 2025, even as it accelerates model releases to compete with US players such as OpenAI and domestic rival DeepSeek. Investor sentiment, however, has been more constructive on the monetization trajectory, with shares surging as much as 35% following recent disclosures and remaining elevated, including gains of more than 18% in subsequent trading. The company has already raised prices multiple times this year, including a 30% increase for its coding plan in February, which could indicate a broader shift across China's AI sector toward prioritizing revenue generation after years of heavy spending.
