Latest Content
GLM 5.1 FP8 icon

GLM 5.1 FP8

NVIDIA
GLM 5.1 is a large-scale Mixture-of-Experts (MoE) language model and a successor to GLM-5 with improved capabilities and performance, designed for long-context reasoning, coding, and agentic workflows. The model has 744B total parameters with 40B activated, using a 78-layer transformer with 6,144 hidden size and 64 attention heads. It routes tokens across 256 experts with top-8 selection for efficient sparse computation. The model supports up to ~202K context length with RoPE scaling and uses FP8 (e4m3) quantization to reduce memory and improve throughput while maintaining strong multi-step reasoning performance.
TypeMoE LLM
CapabilitiesText Generation, Instruction Following, Reasoning, Mathematical Reasoning+5 more
Group Release DateApril 6, 2026
Links
Blog|HF Model Card
LicenseMIT

Inference Instructions

Deploy and run this model on NVIDIA B200 GPUs using the command below. Copy the command to get started with inference.

CONSOLE
docker run -it --rm 
 --runtime=nvidia 
 --gpus all 
 --ipc=host 
 --shm-size=128g 
 -p 8000:8000 
 -v ~/.cache/huggingface:/root/.cache/huggingface 
 -e HF_TOKEN='YOUR_HF_TOKEN' 
 vllm/vllm-openai:glm51-cu130 
 zai-org/GLM-5.1-FP8 
  --tensor-parallel-size 8 
 --max-model-len auto 
  --gpu-memory-utilization 0.90 
 --tool-call-parser glm47 
 --reasoning-parser glm45 
 --enable-auto-tool-choice 
 --max-num-seqs 1024 
 --trust-remote-code
Note

For serving this model, it is recommended to use the vllm/vllm-openai:glm51-cu130 docker image for CUDA 13.0+ or an image with vLLM v0.19.0+

Model Benchmarks

Each model was tested with a fixed input size and total token volume while increasing concurrency to measure serving performance under load.

ITL vs Concurrency

Time to First Token

Throughput Scaling

Total Tokens/sec vs Avg TTFT

Vultr Cloud GPU

NVIDIA HGX B200

Deploy NVIDIA B200 on Vultr Cloud GPU

Deploy NVIDIA B200 instancesContact Sales