Zhipu Launches First Native Model for Lobster Scenarios GLM-5-Turbo, Raises API Prices As Well

On March 16, Zhipu AI launched GLM-5-Turbo, the world's first general-purpose foundation model natively optimized for lobster scenarios. It supports up to 128K output tokens and a 200K context window, with enhanced tool calling, command tracking, persistent task handling, and long-chain execution capabilities. It also integrates reasoning mode, streaming output, function calling, context caching, and MCP. 

Upon release, the company simultaneously increased API prices for GLM-5-Turbo by 20%. The model is currently in an experimental, closed-source phase, with its research outcomes and features slated for integration into Zhipu's next open-source model. 

Notably, alongside the launch of this "lobster-dedicated brain" model, Zhipu also introduced lobster-themed packages for both individual and enterprise users. A monthly experience plan is available for just RMB 39 yuan, offering 35 million tokens, while an advanced monthly plan is priced at RMB 99 yuan for 100 million tokens. This strategy aims to empower both businesses and individuals to easily achieve "token freedom" in lobster scenarios, lowering the barrier to entry for developers.