willitrun·ai

Can Qwen 3 32B run on RTX 3090 Ti 24GB?

BARELY — Tight on Memory

A80Great
Estimated from fit model

Qwen 3 32B needs ~27.0 GB VRAM. RTX 3090 Ti 24GB has 24.0 GB. With Q4_K_M quantization, expect ~23 tok/s.

Runtime: OllamaCapacity: OffloadBandwidth: HighStack: BasicBottleneck: 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.

Capabilities:

Select quantization to explore

Q4_K_M (Medium quality) — 27.0 GB, 23.3 tok/s, Very compromised (needs ~2.2 GB host RAM)
27.0 GB required24.0 GB available
113% VRAM needed

3.0 GB over capacity — needs offload or smaller quantization

Fit status

Very compromised (needs ~2.2 GB host RAM)

Decode

23.3 tok/s

TTFT

8313 ms

Safe context

4K

Memory

27.0 GB / 24.0 GB

Offload

10%

Memory breakdown

Weights19.5 GB
KV Cache3.9 GB
Runtime1.2 GB
Headroom2.4 GB

See how fast it feels

See how fast it feelsQwen 3 32B on RTX 3090 Ti 24GB
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: 23.3 tok/s decode · 8.3s TTFT (warm) · 58 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 2.2 GB of extra host RAM just for the offloaded portion, before OS and other tools.

Performance by workload

WorkloadGradeFitDecodeTTFTContext
ChatSRuns with offload (needs ~0.8 GB host RAM)27.3 tok/s3872 ms4K
CodingAVery compromised (needs ~2.2 GB host RAM)23.3 tok/s8313 ms4K
Agentic CodingFToo heavy17.5 tok/s16065 ms4K
ReasoningAVery compromised (needs ~2.2 GB host RAM)23.3 tok/s9824 ms4K
RAGFToo heavy17.5 tok/s20081 ms4K

Inference speed

Qwen 3 32B inference speed — tokens per second by GPU & Mac

Estimated decode speed (tokens/sec) for Qwen 3 32B at Q4_K_M across popular GPUs and Apple Silicon, using the fastest local runtime per device. Fastest is RTX 5090 32GB at ~67 tok/s. Speed is memory-bandwidth bound, so cards that fit the whole model in VRAM run far faster than ones that offload to system RAM.

GPU / MacMemoryQuantSpeed (tok/s)Fits?
NVIDIARTX 5090 32GB
32 GBQ4_K_M66.9Tight
MacBook Pro M4 Max 128GB
128 GBQ4_K_M33.5Fits
MacBook Pro M4 Max 64GB
64 GBQ4_K_M33.5Fits
Mac Studio M3 Ultra 256GB
256 GBQ4_K_M31.0Fits
Mac Studio M2 Ultra 128GB
128 GBQ4_K_M25.9Fits
NVIDIARTX 4090 24GB
24 GBQ4_K_M24.9Heavy offload
Mac Studio M1 Ultra 128GB
128 GBQ4_K_M24.5Fits
RX 7900 XTX 24GB
24 GBQ4_K_M23.0Heavy offload
NVIDIARTX 3090 24GB
24 GBQ4_K_M21.3Heavy offload
MacBook Pro M4 Pro 48GB
48 GBQ4_K_M21.1Tight
MacBook Pro M3 Max 64GB
64 GBQ4_K_M13.4Fits
MacBook Pro M1 Max 64GB
64 GBQ4_K_M12.3Fits
NVIDIARTX 4080 Super 16GB
16 GBQ4_K_M9.0Too big
NVIDIARTX 4070 12GB
12 GBQ4_K_M3.2Too big
NVIDIARTX 3060 12GB
12 GBQ4_K_M2.0Too big
NVIDIARTX 4060 8GB
8 GBQ4_K_M2.0Too big

Estimates for single-stream decoding at Q4_K_M; real tokens/sec varies with prompt length, context, batch size, and runtime build. Prompt processing (prefill) is faster than the decode figures shown here.

Quantization options

How Qwen 3 32B (32B params) fits at each quantization level on RTX 3090 Ti 24GB (24.0 GB usable).

QuantBitsVRAMQualityFit
Q2_K
2
12.5 GB
LowS91
Q3_K_S
3
15.7 GB
LowS91
NVFP4Best for your GPU
4
17.9 GB
MediumS90
Q4_K_M
4
19.5 GB
MediumF0
Q5_K_M
5
23.0 GB
HighF0
Q6_K
6
26.2 GB
HighF0
Q8_0
8
34.2 GB
Very HighF0
F16
16
65.6 GB
MaximumF0

Get started

Copy-paste commands to run Qwen 3 32B on your machine.

Run

ollama run qwen3:32b

Your hardware

More models your RTX 3090 Ti 24GB can run

ModelParamsGradeDecodeCapabilities
AlibabaQwen 3.5 35B A3B35BA60.6 tok/s

Frequently asked questions

Can RTX 3090 Ti 24GB run Qwen 3 32B?

Yes, RTX 3090 Ti 24GB can run Qwen 3 32B with a A grade (Very compromised (needs ~2.2 GB host RAM)). Expected decode speed: 23.3 tok/s.

How much VRAM does Qwen 3 32B need?

Qwen 3 32B (32B parameters) requires approximately 27.0 GB of memory with Q4_K_M quantization.

What is the best quantization for Qwen 3 32B?

The recommended quantization for Qwen 3 32B is Q4_K_M, which balances quality and memory efficiency.

What speed will Qwen 3 32B run at on RTX 3090 Ti 24GB?

On RTX 3090 Ti 24GB, Qwen 3 32B achieves approximately 23.3 tokens per second decode speed with a time-to-first-token of 8313ms using Q4_K_M quantization.

Can RTX 3090 Ti 24GB run Qwen 3 32B for coding?

For coding workloads, Qwen 3 32B on RTX 3090 Ti 24GB receives a A grade with 23.3 tok/s and 4K context.

What context window can Qwen 3 32B use on RTX 3090 Ti 24GB?

On RTX 3090 Ti 24GB, Qwen 3 32B can safely use up to 4K tokens of context. The model's official context limit is 131K, but available memory constrains the safe maximum.

What should I upgrade first if Qwen 3 32B feels slow on RTX 3090 Ti 24GB?

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.

See all results for RTX 3090 Ti 24GBSee all hardware for Qwen 3 32B
Embed this result▼

Paste this snippet into any page to show a live fit card.

<iframe src="https://willitrunai.com/embed/qwen-3-32b-on-rtx-3090-ti-24gb" width="400" height="180" frameborder="0" style="border:none;border-radius:12px;overflow:hidden;" title="Will It Run AI — fit result"></iframe>

Preview: