nim

Invoke various LLMs (GLM-5, Kimi-k2.5, Llama 3.1, etc.) via NVIDIA NIM API to save main agent tokens and leverage specialized model capabilities.