VOOZH about

URL: https://willitrunai.com/can-run/qwen-3-235b-a22b-on-h20-96gb

⇱ Qwen 3 235B A22B on NVIDIA H20 96GB? No — Alternatives

Can Qwen 3 235B A22B run on NVIDIA H20 96GB?

YES — With Q2_K

A81Great
○Estimated from fit model

Qwen 3 235B A22B needs ~105.0 GB VRAM. NVIDIA H20 96GB has 96.0 GB. With Q2_K quantization, expect ~65 tok/s.

Runtime: llama.cppCapacity: OffloadBandwidth: HighStack: StandardBottleneck: Host offload
Share:

Operating mode

Choose the run profile you care about

Interactive favors responsiveness, while light API and scale-out lean harder on serving readiness. The fit stays the same, but the recommendation lens changes.

Current mode

Balanced

Balanced for general local use. Keeps the ranking neutral across personal and serving workflows.

Qwen 3 235B A22B at Q4_K_M needs 156.7 GB — too much for NVIDIA H20 96GB (96.0 GB). Runs at Q2_K (105.0 GB) with low quality.
Capabilities:

Select quantization to explore

Q4_K_M (Medium quality) — 156.7 GB, exceeds 96.0 GB available
156.7 GB required96.0 GB available
163% VRAM needed

60.7 GB over capacity — needs offload or smaller quantization

Fit status

Too heavy

Decode

25.1 tok/s

TTFT

7727 ms

Safe context

4K

Memory

156.7 GB / 96.0 GB

Offload

40%

Memory breakdown

Weights143.4 GB
KV Cache2.9 GB
Runtime0.9 GB
Headroom9.6 GB

See how fast it feels

With memory offload — actual speed may be lower
See how fast it feelsQwen 3 235B A22B on NVIDIA H20 96GB
1st promptCold start — includes initialization
>Solve: If a train travels at 60 mph for 2.5 hours, how far does it go?
Let me work through this step by step. Given information: - Speed = 60 mph (miles per hour) - Time = 2.5 hours Formula: Distance = Speed x Time Calculation: Distance = 60 mph x 2.5 h = 150 miles The train travels 150 miles. Double-check: at 60 mph the train covers 60 miles each hour. In half an hour it covers 30 miles. So in 2.5 hours: 60 + 60 + 30 = 150 miles. Confirmed.
2nd promptWarm — model ready, no init overhead
>What if the train stops for 20 minutes halfway?
The stop doesn't change the total distance covered while moving, but it affects total trip time. Breakdown: - First half: 1.25 hours at 60 mph = 75 miles - Stop: 20 minutes (no distance) - Second half: 1.25 hours at 60 mph = 75 miles Total distance: 150 miles (same) Total time: 2.5 h + 0.33 h = 2.83 hours Average speed: 150 / 2.83 = 53 mph The distance stays the same but average speed drops to 53 mph because of the stop.
Estimated: 25.1 tok/s decode · 7.7s TTFT (warm) · 63 tok/s prefill

What limits this setup

It fits through host-memory offload, and offload is the main reason performance drops.

CPU or host-memory offload is active

About 10% of the working set spills out of accelerator memory, which usually hurts latency and sustained decode throughput.

Very little memory headroom

You can run the model, but there is not much room left for longer context, bigger batches, extra apps, or future model updates.

Best improvement path

Remove offload with more accelerator memory

Prioritize a GPU or unified-memory tier that fits the whole model natively. Removing offload usually helps more than small compute gains.

Buy headroom, not only minimum fit

A slightly larger memory tier gives you safer context growth and makes the recommendation more future-proof.

Increase host RAM if you keep offloading

This setup may need roughly 7.9 GB of extra host RAM just for the offloaded portion, before OS and other tools.

Performance by workload

WorkloadGradeFitDecodeTTFTContext
ChatFToo heavy25.4 tok/s4151 ms4K
CodingFToo heavy25.1 tok/s7727 ms4K
Agentic CodingFToo heavy24.3 tok/s11581 ms4K
ReasoningFToo heavy25.1 tok/s9132 ms4K
RAGFToo heavy24.3 tok/s14477 ms4K

Quantization options

How Qwen 3 235B A22B (235B params) fits at each quantization level on NVIDIA H20 96GB (96.0 GB usable).

QuantBitsVRAMQualityFit
Q2_K
2
91.7 GB
LowF0
Q3_K_S
3
115.2 GB
LowF0
NVFP4
4
131.6 GB
MediumF0
Q4_K_M
4
143.4 GB
MediumF0
Q5_K_M
5
169.2 GB
HighF0
Q6_K
6
192.7 GB
HighF0
Q8_0
8
251.5 GB
Very HighF0
F16
16
481.7 GB
MaximumF0

Get started

Copy-paste commands to run Qwen 3 235B A22B on your machine.

Run

lms load Qwen3-235B-A22B-Instruct-2507 && lms server start

Upgrade options

Hardware that runs Qwen 3 235B A22B well

👁 NVIDIA
NVIDIA B200 180GBBudget pick
180 GB VRAM (+84)8000 GB/s (+4000)
S
Makes the model fit on the accelerator instead of staying completely out of reach.136.8 tok/s decode

Makes the model fit on the accelerator instead of staying completely out of reach.

Removes host-memory offload, which is usually the single biggest latency and throughput win.

~$30,000 MSRP

👁 NVIDIA
NVIDIA H200 141GBBest value
141 GB VRAM (+45)4800 GB/s (+800)
A
Makes the model fit on the accelerator instead of staying completely out of reach.56.1 tok/s decode

Makes the model fit on the accelerator instead of staying completely out of reach.

Raises estimated decode speed by about 124%.

~$30,000 MSRP

👁 NVIDIA
NVIDIA H200 PCIe 141GBNVIDIA upgrade
141 GB VRAM (+45)4800 GB/s (+800)
A
Makes the model fit on the accelerator instead of staying completely out of reach.47.7 tok/s decode

Makes the model fit on the accelerator instead of staying completely out of reach.

Raises estimated decode speed by about 90%.

~$30,000 MSRP

Frequently asked questions

See all results for NVIDIA H20 96GBSee all hardware for Qwen 3 235B A22B