Make every AI credit count: See, cap, and reduce agent spend

This post introduces new capabilities for controlling AI agent spend within the GitLab Duo Agent Platform. It details the use of GitLab-managed open-weight models (GLM 5.3, Kimi K3, MiniMax M3) to match models to tasks and reduce costs, the integration of GitLab Orbit for providing lifecycle context to agents to cut down on retries, and the introduction of Duo Agent Platform Impact Analytics and the GitLab Credits dashboard for tracing spend by user, team, task, and model. It also highlights credit and usage controls for setting caps at various levels to manage consumption.