Qwen3-32B TPS calculator

Open weights Alibaba 32.8B parameters April 2025

Each card below is assessed against this model at the context length and minimum quality you choose. Speed is an estimate for a single request, calculated from the card's memory bandwidth and the size of the model once compressed.

Calculated for this model

132 cards that can run it

818 cards we hold specifications for

Smallest card that fits

RTX A4500

20 GB · Q3_K_M · 22.3 tok/s

Fastest card

B200

103 tok/s · 180 GB

Which GPUs can run Qwen3-32B?

Set the inputs, read the answer

Context length tokens

A longer conversation needs more memory, which can push this model off smaller cards.

Minimum quality

Hides cards that would only fit the model by compressing it below this point.

Search cards

Card memory

Also show cards that cannot run it

132 cards match

Calculating

					Needs	Quantisation	Fit
103 tok/s 88–124	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	35.8 GB	Q8_0	Comfortable
103 tok/s 88–124	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	35.8 GB	Q8_0	Comfortable
82.5 tok/s 49–132 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	35.8 GB	Q8_0	Comfortable
82.5 tok/s 49–132 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	35.8 GB	Q8_0	Comfortable
66.0 tok/s 40–106 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	35.8 GB	Q8_0	Comfortable
63.1 tok/s 54–76	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	35.8 GB	Q8_0	Comfortable
63.1 tok/s 54–76	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	35.8 GB	Q8_0	Comfortable
60.4 tok/s 36–97 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	35.8 GB	Q8_0	Comfortable
53.6 tok/s 32–86 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	35.8 GB	Q8_0	Comfortable
53.6 tok/s 32–86 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	35.8 GB	Q8_0	Comfortable
53.6 tok/s 32–86 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	35.8 GB	Q8_0	Comfortable
50.9 tok/s 43–61	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	35.8 GB	Q8_0	Comfortable
43.4 tok/s 37–52	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	35.8 GB	Q8_0	Comfortable
43.4 tok/s 37–52	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	35.8 GB	Q8_0	Comfortable
43.4 tok/s 37–52	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	35.8 GB	Q8_0	Comfortable
43.4 tok/s 37–52	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	35.8 GB	Q8_0	Comfortable
43.4 tok/s 37–52	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	35.8 GB	Q8_0	Comfortable
40.0 tok/s 34–48	GeForce RTX 5090 D V2 NVIDIA	24 GB	1,340 GB/s	Aug 2025	20.5 GB	Q4_K_M	Tight
36.4 tok/s 31–44	A30X NVIDIA	24 GB	1,220 GB/s	Apr 2021	20.5 GB	Q4_K_M	Tight
35.1 tok/s 30–42	DRIVE A100 PROD NVIDIA	32 GB	1,870 GB/s	May 2020	28.1 GB	Q6_K	Tight
35.1 tok/s 30–42	GRID A100A NVIDIA	32 GB	1,870 GB/s	May 2020	28.1 GB	Q6_K	Tight
33.6 tok/s 29–40	GeForce RTX 5090 NVIDIA	32 GB	1,790 GB/s	Jan 2025	28.1 GB	Q6_K	Tight
33.6 tok/s 29–40	GeForce RTX 5090 D NVIDIA	32 GB	1,790 GB/s	Jan 2025	28.1 GB	Q6_K	Tight
33.0 tok/s 20–53 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	35.8 GB	Q8_0	Comfortable
33.0 tok/s 20–53 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	35.8 GB	Q8_0	Comfortable

Speeds are estimates for a single request — one conversation at a time — calculated from memory bandwidth, model size and quantisation. Real throughput varies with the inference runtime and its version. Figures published by hardware vendors measure many simultaneous requests and are much higher.

On record

Full specification

Everything on record for this model. Most of it describes how it was trained rather than how it runs — useful context for judging how much work went into it, and how it compares with models built at a different scale.

Origin

Who built this model, where, and when it was published.

Organisation: Alibaba
Organisation type: Industry
Country: China
Published: 29 April 2025

What it does

The problem areas the model was built for. A model can carry several of each.

Domain: Language
Task: Language modeling/generation, Question answering, Mathematical reasoning, Quantitative reasoning, Code generation, Translation

Size

How large the model is and how much data it was trained on. Parameters are the figure that decides whether it fits on a given graphics card.

Parameters: 32.8B
Training data: tokens
Epochs: 1

Training compute

The arithmetic performed to train the model, measured in floating-point operations. It is a measure of what the training run cost, not of how fast the finished model answers you.

Training compute: 7.1 × 10²⁴ FLOP
How it was established: Operation counting

Availability

Whether you can obtain the model and run it on your own hardware, which is what decides if any of the graphics-card figures on this page apply.

Weights: Open — downloadable
Model access: Open weights (unrestricted)
Training code: Unreleased
Hugging Face: Qwen

How it is classified

Labels the source dataset applies when tracking notable models, and how confident it is in the entry.

Likely above 10²³ FLOP: Yes
Record confidence: Confident

Sources

Where this record came from and when it was last checked.

Reference: Qwen3: Think Deeper, Act Faster
Last updated: 28 November 2025

The extremes

The ten fastest GPUs that run Qwen3-32B

Ranked by estimated tokens per second, newest card first where speeds tie. Because generation is bound by memory bandwidth, this ordering follows bandwidth rather than any gaming benchmark.

The smallest GPUs that still run Qwen3-32B

The cheapest route in, by memory capacity. A tight fit runs the model but leaves nothing spare for a longer conversation.

What the numbers mean

Hardware requirements in practice

Minimum card

RTX A4500

Memory needed

16.7 GB

Fastest

103 tok/s

With 32.8B parameters, Qwen3-32B lands in the range a serious desktop card can handle once the weights are compressed. 132 of the cards we track can run it.

The entry point is the RTX A4500: 20 GB of memory, Q3_K_M compression, roughly 22.3 tokens per second.

The quickest result comes from a B200 at around 103 tokens per second — its 8,000 GB/s of bandwidth is what buys that.

Background

Qwen3-32B was published by Alibaba, in China, in April 2025. It comes out of industry.

It works in Language, and is recorded as doing language modeling/generation, Question answering, Mathematical reasoning, Quantitative reasoning, Code generation, Translation.

Published weights mean the model runs on your machine rather than someone else's, which is what makes the hardware question below answerable at all. It is published under the Qwen organisation on Hugging Face.

Reading the throughput figures

Across every card that can run it, the middle of the range is about 20.1 tokens per second, and 101 of them clear the ten tokens per second that roughly matches reading speed.

Being dense, it reads all of itself per token, which is why the ordering by speed below follows the ordering by memory bandwidth so closely.

Because the architecture is recorded, the memory column is derived rather than estimated.

How it was trained

The training run consumed about 7.1 × 10²⁴ FLOP. That figure describes the cost of creating it and has no bearing on how quickly it generates text.

Step by step

How to choose a GPU for Qwen3-32B

The table above has already assessed every card we hold specifications for against this model. Getting to your answer takes six steps.

01

Start from the memory column

The table lists every card that can hold Qwen3-32B — around 16.7 GB at Q3_K_M. That figure, not the card's headline performance, is what decides whether it runs.
02

Set the context length you will work at

Longer conversations cost memory on top of what the weights need. Move the slider to your real working length before trusting any row for Qwen3-32B.
03

Decide how much compression you will accept

The quantisation column varies by card, because a bigger card holds a more accurate copy of Qwen3-32B — Q3_K_M on the smallest card that fits. Set a floor to hold the comparison at one level.
04

Rank by throughput rather than spec sheet

The speed ordering for Qwen3-32B is effectively an ordering by memory bandwidth, which is why the B200 tops it at 103 tok/s.
05

Check the fit verdict before buying

The fit column separates cards that just manage Qwen3-32B from those with room to spare. Buy for the second if the context might grow.
06

See what else that card runs

Following a card through to its own page shows every other model it can hold, which is the question that follows once Qwen3-32B is settled.

Answers

Qwen3-32B — common questions

How many parameters does Qwen3-32B have?

Qwen3-32B has 32.8B parameters. Number of Parameters: 32.8B Number of Paramaters (Non-Embedding): 31.2B Number of Layers: 64 Number of Attention Heads (GQA): 64 for Q and 8 for KV Context Length: 32,768 natively and 131,072 tokens with YaRN. That figure is the total, and it is what decides how much memory the model needs — roughly half a gigabyte per billion at the compression most people use.

Who created Qwen3-32B?

Qwen3-32B was published by Alibaba, based in China, categorised as industry.

When was Qwen3-32B released?

Qwen3-32B was published in April 2025.

What is Qwen3-32B used for?

Qwen3-32B works in Language, and is recorded as handling language modeling/generation, Question answering, Mathematical reasoning, Quantitative reasoning, Code generation, Translation. These are the areas it was designed around; they describe intent rather than a hard boundary.

Where can I download Qwen3-32B?

Its weights are published under the Qwen organisation on Hugging Face. We do not host model files — this site calculates what hardware is needed to run them.

How much compute was used to train Qwen3-32B?

Around 7.1 × 10²⁴ FLOP. That measures what producing the model cost and says nothing about how quickly it answers once trained — inference speed comes from memory bandwidth, not from the training budget.

Can I run Qwen3-32B if it does not fit in my GPU?

Partly. Layers that do not fit sit in system memory and run at a fraction of the speed, so a mostly-offloaded Qwen3-32B is rarely worth using — the nearest miss we calculate is short by 6.1 GB. Every figure here assumes the whole model is on the card.

Would two GPUs run Qwen3-32B faster?

A second card roughly doubles the memory available but not the generation rate. With 132 cards already able to run Qwen3-32B alone, the case for pairing is weak.

Why does the quantisation differ between cards for Qwen3-32B?

Because capacity varies, so does how hard Qwen3-32B has to be squeezed — 5 distinct levels appear in the table above. Set a minimum quality to compare at one.

How accurate are these Qwen3-32B speed estimates?

They are calculated from specifications rather than measured, and each carries a range — 88–124 tok/s on the B200, for instance. The same model and card vary by thirty to fifty per cent depending on the inference software and its version.

What GPU do I need to run Qwen3-32B?

The smallest card in our catalogue that holds Qwen3-32B is the RTX A4500, with 20 GB of memory. It runs the model at Q3_K_M using about 16.7 GB, and produces roughly 22.3 tokens per second. 132 cards in total can run it.

How fast is Qwen3-32B on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 103 tokens per second; the slowest that still runs it manages considerably less. Reading speed is around ten tokens per second, and 101 of the cards that can run Qwen3-32B clear that.

How much VRAM does Qwen3-32B need?

About 16.7 GB at Q3_K_M compression, which is what the smallest card that runs it uses. Less compression needs more: the figures in the memory column above are recalculated for each card, because each one holds the least-compressed version it can.

Can I run Qwen3-32B on a 24 GB GPU?

Yes. A GeForce RTX 5090 D V2 with 24 GB runs it at Q4_K_M, using about 20.5 GB and generating roughly 40.0 tokens per second — a tight fit.

Is Qwen3-32B open source?

Its weights are published, so Qwen3-32B can be downloaded and run on your own hardware. Note that open weights is not the same as open source in the full sense — it says nothing about the training data, the training code, or the commercial terms attached.

Source

Original publication

Record last updated 28 November 2025

The other direction

Looking at it from the other side?

This page starts from the model. If you already own a card and want to know everything it will run, start from the hardware instead.

All AI models