QriticalBeta
Open Model Catalogue
Z.ai
Z.ai
Text generationReasoningCoding

GLM-5

zai-org/GLM-5-FP8

Z.ai flagship — MIT, 8 × H200

Deploy GLM-5 nowView on HuggingFace
Parameters
~745B MoE · FP8
Min VRAM
≈ 830 GB
Hardware
11 × 80 GB GPU
OpenAI-compatible API with a per-workload key — drop-in for existing SDKs.
Dedicated GPUs — your model, your instance, no shared queues.
Billed by the second with auto-shutdown — never pay for idle time you didn't plan.