Latest Content
Llama 3.1 405B Instruct icon

Llama 3.1 405B Instruct

NVIDIA
Llama-3.1-405B-Instruct is an instruction-tuned dense transformer large language model designed for advanced reasoning, multilingual dialogue, and large-scale AI applications. Built on a 405B parameter architecture, it features 126 transformer layers, 128 attention heads, and a 16,384 hidden size, utilizing Grouped Query Attention (GQA) for efficient large-scale inference. The model supports up to a 128K token context window with RoPE scaling for long-context understanding. Fine-tuned using supervised learning and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety.
TypeDense LLM
CapabilitiesText Generation, Instruction Following, Reasoning, Mathematical Reasoning+5 more
Group Release DateJuly 22, 2024
Links
Blog|HF Model Card
LicenseLlama3.1

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' 
 -e LD_LIBRARY_PATH='/usr/local/nvidia/lib64:/usr/local/nvidia/lib:/usr/lib/x86_64-linux-gnu' 
 vllm/vllm-openai:v0.15.0-cu130 
 meta-llama/Llama-3.1-405B-Instruct 
  --tensor-parallel-size 8 
 --max-model-len auto 
  --max-num-batched-tokens 65536 
  --gpu-memory-utilization 0.95 
  --quantization fp8 
 --kv-cache-dtype fp8 
 --max-num-seqs 1024 
 --disable-log-requests 
 --trust-remote-code

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