Latest Content
Qwen3.5 397B A17B icon

Qwen3.5 397B A17B

NVIDIA
Qwen3.5-397B-A17B is a multimodal Mixture-of-Experts (MoE) large language model optimized for long-context reasoning, vision-language understanding, and scalable agentic workflows. It contains 397B total parameters with 17B active parameters, activating 10 routed experts plus 1 shared expert from 512 experts per token. The model employs a 60-layer hybrid architecture combining Gated DeltaNet and full attention layers, with 32 attention heads and a 4,096 hidden size. Supporting a 262K token context window, extendable to ~1M, it integrates a 27-layer vision encoder for unified multimodal processing.
TypeVision-Language Model
CapabilitiesText Generation, Instruction Following, Reasoning, Mathematical Reasoning+6 more
Paper/Blog
Blog|HF Model Card
LicenseApache 2.0

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 --gpus all 
 --shm-size 128g 
 -p 8000:8000 
 -v ~/.cache/huggingface:/root/.cache/huggingface 
 -e HF_TOKEN='YOUR_HF_TOKEN' 
 --ipc=host 
 lmsysorg/sglang:v0.5.9 
 python3 -m sglang.launch_server 
 --model-path Qwen/Qwen3.5-397B-A17B 
 --host 0.0.0.0 
 --port 8000 
 --max-running-requests 1024 
 --max-prefill-tokens 65536 
 --tp 8 
 --ep 8 
 --tool-call-parser qwen3_coder 
 --reasoning-parser qwen3 
 --mem-fraction-static 0.95 
 --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