Model Library/Qwen3-235B-A22B

Qwen

Qwen3-235B-A22B

A 235B mixture-of-experts model. Two linked Sparks over llama.cpp RPC.

TextvLLM1 builds

Recommended: B200, 180 GB HBM3e, $5.49 / hour, billed per minute. Check availability.

Configure

Every build below is published on Hugging Face and fits at least one GPU we rent. Pick the quantization and engine: the memory it needs, the GPUs it fits and the recommended GPU follow, and Deploy takes that exact build, revision and GPU to the console. Only the one-click preset is tested by GPUwerk.

Qwen3-235B-A22B, NVFP4 on vLLM
QuantizationNVFP4
EnginevLLM
StatusUntested by GPUwerk
Weights134.1 GB
GPU memory it needsAbout 157.2 GB with a 32k-token cache, an estimate
Recommended GPUB200, $5.49 / hour
Weights sourcenvidia/Qwen3-235B-A22B-NVFP4, revision 21cfa2c9e152
Licence on the cardapache-2.0
GPUs Qwen3-235B-A22B NVFP4 does and does not fit
GPUMemoryPriceAvailabilityDeploy
RTX 5090x86 · Blackwell32 GB GDDR7Does not fit: needs about 157 GB of GPU memory (weights 134.1 GB plus cache, an estimate); RTX 5090 offers 29 GB to it.
DGX SparkArm · GB10128 GB unifiedDoes not fit: needs about 157 GB of GPU memory (weights 134.1 GB plus cache, an estimate); DGX Spark offers 96 GB to it.
RTX PRO 6000x86 · Blackwell96 GB GDDR7Does not fit: needs about 157 GB of GPU memory (weights 134.1 GB plus cache, an estimate); RTX PRO 6000 offers 86 GB to it.
H100 SXMx86 · Hopper80 GB HBM3Does not fit: NVFP4 weights need a Blackwell GPU; H100 is not one.
H200x86 · Hopper141 GB HBM3eDoes not fit: NVFP4 weights need a Blackwell GPU; H200 is not one.
B200x86 · Blackwell · recommended180 GB HBM3e$5.49 / hourCheck availabilityRun it yourself Qwen3-235B-A22B NVFP4 on B200
B300x86 · Blackwell288 GB HBM3e$6.49 / hourCheck availabilityRun it yourself Qwen3-235B-A22B NVFP4 on B300

Builds by the model's maker, NVIDIA and Unsloth, read from Hugging Face on 21 November 2026. Memory is an estimate (weights x 1.15 plus 3 GB); GGUF runs on llama.cpp, which GPUwerk builds for the DGX Spark only, and NVFP4 needs a Blackwell GPU. SGLang is not offered: there is no SGLang runtime yet.

Details

About the model.

Qwen3-235B-A22B
MakerQwen
TaskText
Parameters235B total, 22B active (mixture of experts)
LicenceSee the model card

No speed figure here: GPUwerk lists only figures it measured itself. Reported figures, with their sources, are in the model guide.

Quick start

  1. Deploy a 2 × DGX Spark with a base image (DGX OS or Ubuntu).
  2. Follow the Qwen3-235B-A22B guide to download the weights and start llama.cpp.
  3. Connect over SSH or the browser terminal; the machine is yours until you terminate it.

SSH, the browser terminal, VNC and your workspace work the same on every instance. Terminate when you are done and billing ends.

Related models

QwenCoding · Text

Qwen3 Coder 30B

A coding model for agents and editors, with tool calls turned on.

One-click preset · vLLM · RTX 5090

QwenText

Qwen3.6 35B-A3B

A fast mixture-of-experts model for chat and agents, also on the pay-per-token API.

One-click preset · llama.cpp · DGX Spark

OpenAIText

GPT-OSS 120B

OpenAI's open-weight 120B model, quantized to fit one DGX Spark.

One-click preset · llama.cpp · DGX Spark

All models in the library