Qwen3 235B-A22B (MoE): Hardware & Business Fit
- Tools
- Reasoning
- Code
- Multilingual
- Long context
A frontier-class open MoE. Memory is bounded by total params; throughput benefits from sparse activation. Figures are placeholders — verify before planning hardware.
- Parameters
- ~235B (≈22B active, MoE)
- Context
- ~128K tokens
- Deployment
- cloud
- VRAM @ 4-bit
- ~130GB
What Qwen3 235B-A22B (MoE) is good for
- ▸Frontier-quality private serving
- ▸Org-wide platform
Best quantization choices
Approximate memory per quantization (weights + KV cache at modest context). Treat as ±.
| Quant | ~Memory | When to use |
|---|---|---|
| Q4_K_M | ~130GB | Best size/quality trade-off — the usual default for local serving. |
| Q8_0 | ~235GB | Higher fidelity; ~1.7× the memory of 4-bit. |
| FP16 | ~470GB | Full precision; largest footprint, best quality. |
Run Qwen3 235B-A22B (MoE) locally
Pull and run with Ollama, or grab the weights from Hugging Face.
$ ollama run qwen3:235bQwen/Qwen3-235B-A22BCompatible hardware
Devices from our catalog graded for Qwen3 235B-A22B (MoE), best fit first.
- NVIDIA B200 (placeholder)NVIDIA · Datacenter GPUs
Fits at Q4_K_M (~130GB) with ~39GB headroom — about 1 concurrent instance.
Q4_K_M · ~130GBRuns well - Supermicro 8x H100 SuperServerSupermicro · AI Servers
Fits at FP16 (~470GB) with ~93.2GB headroom — about 1 concurrent instance.
FP16 · ~470GBRuns well - Dell PowerEdge XE9680Dell · AI Servers
Fits at FP16 (~470GB) with ~93.2GB headroom — about 1 concurrent instance.
FP16 · ~470GBRuns well - AMD Instinct MI300XAMD · Datacenter GPUs
Fits at Q4_K_M (~130GB) with ~39GB headroom — about 1 concurrent instance.
Q4_K_M · ~130GBRuns well - Cloud B200 (Blackwell profile, to verify)Cloud · Cloud GPU Profiles
Fits at Q4_K_M (~130GB) with ~28.4GB headroom — about 1 concurrent instance.
Q4_K_M · ~130GBRuns well - Apple Mac Studio (M2 Ultra)Apple · Apple Silicon
Fits at Q4_K_M (~130GB) but limited bandwidth makes token generation slow for a 235B model.
Q4_K_M · ~130GBRuns slowly - Apple Mac Studio (M4 Ultra class, to verify)Apple · Apple Silicon
Fits at Q4_K_M (~130GB) but limited bandwidth makes token generation slow for a 235B model.
Q4_K_M · ~130GBRuns slowly - NVIDIA H200 (141GB)NVIDIA · Datacenter GPUs
Even the smallest quantization (~130GB) exceeds usable memory (~124.1GB). Choose a smaller model or step up the hardware.
Not recommended - Cloud H200 141GB (profile)Cloud · Cloud GPU Profiles
Even the smallest quantization (~130GB) exceeds usable memory (~124.1GB). Choose a smaller model or step up the hardware.
Not recommended - NVIDIA H100 (80GB)NVIDIA · Datacenter GPUs
Even the smallest quantization (~130GB) exceeds usable memory (~70.4GB). Choose a smaller model or step up the hardware.
Not recommended
Use inside the AI Business OS
Qwen3 235B-A22B (MoE) suits these AI Business OS agent archetypes:
A model is only the engine. Inside the AI Business OS it is wrapped with permissions, tools, connectors, RAG and audit so it can actually do business work safely — see how the AI Business OS works →
Frequently asked questions
What hardware do I need to run Qwen3 235B-A22B (MoE)?+
At 4-bit you need roughly ~130GB of usable memory. The minimum self-hostable option in our catalog is the NVIDIA B200 (placeholder). For a comfortable run we recommend the NVIDIA B200 (placeholder).
Which quantization should I use for Qwen3 235B-A22B (MoE)?+
Q4_K_M is the usual default — the best size/quality trade-off. Step up to Q8_0 or FP16 if you have spare memory and want higher fidelity.
Should I run Qwen3 235B-A22B (MoE) locally or in the cloud?+
Cloud / API is recommended for Qwen3 235B-A22B (MoE). Its size or hosting model makes the cloud the practical choice; pair with smaller local models for everyday private work.
Other sizes in the Qwen family
All Qwen models →Same family, different size. Pick the variant that fits your hardware.
Related models
Similar picks — family siblings and nearest-size models of the same kind.
Use Qwen3 235B-A22B (MoE) inside your AI Business OS
BrainOutput helps you run Qwen3 235B-A22B (MoE) as a private business agent — wrapped with the tools, connectors, RAG and guardrails it needs to do real work on hardware you control.
Use this model in your AI Business OS