Zhipu AI Relaunches GLM Coding Plan with Points-Based Billing
Zhipu AI has relaunched its GLM Coding Plan subscription, introducing a transparent points-based billing system and promotional discounts of up to 30% off until August 15, 2026. The new system converts input, output, cache-hit tokens, and Model Context Protocol (MCP) usage into points to make consumption more predictable. This update standardizes Zhipu AI's pricing model for developer tools, aligning it with modern LLM optimizations like context caching and open standards like MCP. It offers developers more granular control over their AI coding assistant expenses while ensuring predictable billing. Under the new plan, weekend usage is charged at off-peak rates, and users can apply to test a new high-speed version. Existing v1 plan users, who cannot renew their legacy plans, will be allowed to transition to v2 pricing for long-term use.
## BACKGROUND
The Model Context Protocol (MCP) is an open-source standard introduced by Anthropic to unify how LLMs connect with external data sources and tools. Context caching, or prompt caching, is a technique that stores previously processed tokens to reduce latency and costs for repetitive queries in large language models.