GLM 5 Turbo is a foundation model deeply optimized for agentic scenarios. It has been specifically optimized since the training phase for core agent requirements such as tool invocation, instruction following, scheduled and persistent tasks, and long-chain execution. It delivers substantial improvements over GLM-5 in agent workflows, with enhanced stability and reliability in multi-step tasks, stronger comprehension of complex multi-layered instructions, and better time-aware execution for long-running tasks.
Try NowAgentic tool invocation and multi-step execution
Complex instruction decomposition and planning
Persistent long-running agent tasks
200,000 tokens
128,000 tokens
$1.20
$4
$0.24
$15
$0.19