GLM-5.2 is live. #1 throughput on OpenRouter, pay-per-token on FriendliAI. Try it today ➜
NeuralNet-Hub
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
NeuralNet-Hub
Model Tree
Base
Qwen/Qwen3.6-35B-A3B
Quantized
this model
Input Modalities
Text
Image
Video
Output Modalities
Text
Supported Functionality
Dedicated Endpoints
This is an NVFP4-quantized version of Qwen/Qwen3.6-35B-A3B, optimized for deployment on NVIDIA Blackwell architecture GPUs using vLLM.
[!IMPORTANT] NVFP4 quantization requires NVIDIA Blackwell architecture (GB200, RTX 5000 series, etc.). This format is not compatible with Ampere, Ada Lovelace, or Hopper GPUs. If you are running on an older GPU, please use a different quantization format.
Original model: https://huggingface.co/Qwen/Qwen3.6-35B-A3B
This model was quantized to NVFP4 (4-bit NVIDIA Floating Point) using vLLM's built-in quantization pipeline. NVFP4 leverages native FP4 Tensor Core support introduced in Blackwell GPUs, delivering significant memory savings and throughput improvements with minimal quality degradation compared to BF16.
bash
vllm quantize \--model Qwen/Qwen3.6-35B-A3B \--quantization nvfp4 \--output-dir NeuralNet-Hub/Qwen3.6-35B-A3B-NVFP4
This quantized model is intended to be served using vLLM (vllm>=0.9.0 recommended).
bash
vllm serve NeuralNet-Hub/Qwen3.6-35B-A3B-NVFP4 \--quantization nvfp4 \--dtype bfloat16 \--kv-cache-dtype fp8 \--max-model-len 262144 \--reasoning-parser qwen3 \--enable-auto-tool-choice \--tool-call-parser qwen3_coder
yaml
# Deploy with: vllm serve --config config.yaml# Optimized for NVIDIA RTX 6000 PRO (Blackwell)# Benchmarked: ~85-90 parallel requests, up to 1000 tok/sec at higher context lengthsmodel: NeuralNet-Hub/Qwen3.6-35B-A3B-NVFP4dtype: bfloat16kv-cache-dtype: fp8gpu-memory-utilization: 0.95max-model-len: 262144max-num-batched-tokens: 4096max-num-seqs: 200max-cudagraph-capture-size: 209enable-prefix-caching: truetrust-remote-code: truereasoning-parser: qwen3enable-auto-tool-choice: truetool-call-parser: qwen3_coderdefault-chat-template-kwargs: '{"enable_thinking": false}'download-dir: /workspace/modelshost: 0.0.0.0port: 18000
bash
vllm serve --config config.yaml
Qwen3.6 uses a standard chat template compatible with OpenAI-format APIs. Thinking mode is enabled by default.
python
from openai import OpenAIclient = OpenAI(base_url="http://localhost:18000/v1", api_key="EMPTY")messages = [{"role": "user", "content": "Your message here"}]response = client.chat.completions.create(model="NeuralNet-Hub/Qwen3.6-35B-A3B-NVFP4",messages=messages,max_tokens=32768,temperature=1.0,top_p=0.95,extra_body={"top_k": 20},)print(response.choices[0].message.content)
python
response = client.chat.completions.create(model="NeuralNet-Hub/Qwen3.6-35B-A3B-NVFP4",messages=messages,max_tokens=8192,temperature=0.7,top_p=0.8,presence_penalty=1.5,extra_body={"top_k": 20,"chat_template_kwargs": {"enable_thinking": False},},)
python
messages = [{"role": "user","content": [{"type": "image_url", "image_url": {"url": "https://example.com/image.jpg"}},{"type": "text", "text": "Describe this image in detail."}]}]response = client.chat.completions.create(model="NeuralNet-Hub/Qwen3.6-35B-A3B-NVFP4",messages=messages,max_tokens=32768,temperature=1.0,top_p=0.95,extra_body={"top_k": 20},)
| Mode | temperature | top_p | top_k | presence_penalty |
|---|---|---|---|---|
| Thinking — general tasks | 1.0 | 0.95 | 20 | 0.0 |
| Thinking — precise coding | 0.6 | 0.95 | 20 | 0.0 |
| Instruct (non-thinking) | 0.7 | 0.80 | 20 | 1.5 |
bash
pip install -U "huggingface_hub[cli]"
bash
huggingface-cli download NeuralNet-Hub/Qwen3.6-35B-A3B-NVFP4 --local-dir ./Qwen3.6-35B-A3B-NVFP4
bash
huggingface-cli download NeuralNet-Hub/Qwen3.6-35B-A3B-NVFP4 \--include "*.safetensors" \--local-dir ./Qwen3.6-35B-A3B-NVFP4
| Component | Requirement |
|---|---|
| GPU Architecture | NVIDIA Blackwell (sm_100+) |
| VRAM | 24 GB+ recommended |
| CUDA | 12.8+ |
| vLLM | 0.9.0+ |
[!WARNING] NVFP4 is exclusively supported on NVIDIA Blackwell GPUs. Attempting to run this model on Ampere (A100), Ada Lovelace (RTX 4000), or Hopper (H100) will fail. For those architectures, use the original BF16 model or an AWQ/GPTQ quantized variant.
NeuralNet is a pioneering AI solutions provider that empowers businesses to harness the power of artificial intelligence.
Website: https://neuralnet.solutions Email: info[at]neuralnet.solutions