Latest Content
Qwen3.5 35B A3B icon

Qwen3.5 35B A3B

NVIDIA
Qwen3.5-35B-A3B is a multimodal Mixture-of-Experts (MoE) large language model designed for efficient long-context reasoning, multimodal understanding, and scalable deployment. The model contains 35B total parameters with ~3B active parameters, activating 8 routed experts plus 1 shared expert from 256 experts per token. It features a 40-layer hybrid architecture combining Gated DeltaNet and full attention layers, with 16 attention heads and a 2,048 hidden size. The model supports a 262K token context window, extendable to ~1M, and integrates a 27-layer vision encoder for unified multimodal processing.
TypeVision-Language Model
CapabilitiesText Generation, Instruction Following, Reasoning, Mathematical Reasoning+6 more
Group Release DateFebruary 15, 2026
Links
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 -it --rm 
 --runtime=nvidia 
 --gpus all 
 --ipc=host 
 --shm-size=128g 
 -p 8000:8000 
 -v ~/.cache/huggingface:/root/.cache/huggingface 
 -e VLLM_USE_FLASHINFER_MOE_FP16=1 
 -e VLLM_FLASHINFER_MOE_BACKEND=latency 
 -e HF_TOKEN='YOUR_HF_TOKEN' 
 vllm/vllm-openai:v0.17.1 
 Qwen/Qwen3.5-35B-A3B 
  --tensor-parallel-size 8 
 --mm-encoder-tp-mode data 
 --max-model-len auto 
 --max-num-batched-tokens 65536 
 --gpu-memory-utilization 0.95 
 --enable-expert-parallel 
 --tool-call-parser qwen3_coder 
 --reasoning-parser qwen3 
 --enable-auto-tool-choice 
 --max-num-seqs 1024 
 --trust-remote-code
Note

To use VLLM_USE_FLASHINFER_MOE_FP16=1 and VLLM_FLASHINFER_MOE_BACKEND=latency for more optimal performance, need to use expert parallelism.

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