Open SourceModelFree

GLM-5.1

Z.ai's MIT-licensed open-weight MoE (~754B), tuned for long-horizon agentic work across hundreds of tool-call rounds. Server-class; the FP8 checkpoint (~860GB) is the realistic serving target.

Homepage ↗Repo ↗

Alternatives

DeepSeek V4MiniMax M3NVIDIA Nemotron 3 Ultra

1 workflow use GLM-5.1

Local InferenceFreeMachine-verified

Serve GLM-5.1 yourself for long-horizon agentic coding (vLLM)

Stand up the MIT-licensed GLM-5.1 FP8 checkpoint as an OpenAI-compatible endpoint for long agentic runs, validated serve config + endpoint.

vLLMGLM-5.1· advanced