StyleGAN3-R TPS calculator

Open weights NVIDIA,Aalto University 1.6M parameters June 2021

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

818 cards that can run it

818 cards we hold specifications for

Smallest card that fits

Tesla C1080

4 GB · Q8_0 · 23,332 tok/s

Fastest card

B200

2,144,453 tok/s · 180 GB

Which GPUs can run StyleGAN3-R?

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

818 cards match

Calculating

					Needs	Quantisation	Fit
2,144,453 tok/s 1,286,672–3,431,124 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	0.7 GB	Q8_0	Comfortable
2,144,453 tok/s 1,286,672–3,431,124 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	0.7 GB	Q8_0	Comfortable
1,712,399 tok/s 1,027,439–2,739,839 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	0.7 GB	Q8_0	Comfortable
1,712,399 tok/s 1,027,439–2,739,839 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	0.7 GB	Q8_0	Comfortable
1,369,501 tok/s 821,701–2,191,202 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	0.7 GB	Q8_0	Comfortable
1,310,797 tok/s 786,478–2,097,275 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	0.7 GB	Q8_0	Comfortable
1,310,797 tok/s 786,478–2,097,275 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	0.7 GB	Q8_0	Comfortable
1,254,505 tok/s 752,703–2,007,208 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	0.7 GB	Q8_0	Comfortable
1,113,373 tok/s 668,024–1,781,397 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	0.7 GB	Q8_0	Comfortable
1,113,373 tok/s 668,024–1,781,397 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	0.7 GB	Q8_0	Comfortable
1,113,373 tok/s 668,024–1,781,397 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	0.7 GB	Q8_0	Comfortable
1,056,143 tok/s 633,686–1,689,829 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
900,670 tok/s 540,402–1,441,072 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
900,670 tok/s 540,402–1,441,072 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	0.7 GB	Q8_0	Comfortable
900,670 tok/s 540,402–1,441,072 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
900,670 tok/s 540,402–1,441,072 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
900,670 tok/s 540,402–1,441,072 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
685,796 tok/s 411,478–1,097,274 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	0.7 GB	Q8_0	Comfortable
685,796 tok/s 411,478–1,097,274 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	0.7 GB	Q8_0	Comfortable
571,497 tok/s 342,898–914,395 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	0.7 GB	Q8_0	Comfortable
559,300 tok/s 335,580–894,880 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	0.7 GB	Q8_0	Comfortable
546,835 tok/s 328,101–874,937 · low confidence	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	0.7 GB	Q8_0	Comfortable
546,835 tok/s 328,101–874,937 · low confidence	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	0.7 GB	Q8_0	Comfortable
546,835 tok/s 328,101–874,937 · low confidence	A800 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Aug 2022	0.7 GB	Q8_0	Comfortable
546,835 tok/s 328,101–874,937 · low confidence	H100 CNX NVIDIA	80 GB	2,040 GB/s	Mar 2023	0.7 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: NVIDIA,Aalto University
Organisation type: Industry,Academia
Country: United States of America, Finland
Published: 21 June 2021
Authors: Tero Karras, Miika Aittala, Samuli Laine, Erik Härkönen, Janne Hellsten, Jaakko Lehtinen, Timo Aila

What it does

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

Domain: Image generation
Task: Image generation

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: 1.6M
Training data: 50,000,000 tokens
Batch size: 32

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: 2.4 × 10²¹ FLOP
How it was established: Hardware

The training run

What it physically took to train: which chips, how many, for how long, and what that drew from the wall.

Training hardware: NVIDIA V100
Chips used: 8
Wall-clock time: 2,248 hours (93.7 days)
Power draw: 4.9 kW

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 (non-commercial)
Training code: Open (non-commercial)

How it is classified

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

Why it is tracked: Historical significance,Highly cited
Record confidence: Confident
Citations: 2,031

Sources

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

Reference: Alias-Free Generative Adversarial Networks
Last updated: 28 November 2025

The extremes

The ten fastest GPUs that run StyleGAN3-R

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 StyleGAN3-R

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

What you need to run it

Minimum card

Tesla C1080

Memory needed

0.7 GB

Fastest

2,144,453 tok/s

StyleGAN3-R is small enough at 1.6M parameters that hardware is rarely the obstacle — 818 of the cards we track can run it, including cards several years old.

The least hardware that works is a Tesla C1080. Its 4 GB is enough at Q8_0 compression, giving roughly 23,332 tokens per second.

Top of the range is the B200, at roughly 2,144,453 tokens per second thanks to 8,000 GB/s of bandwidth.

Where it came from

StyleGAN3-R was published by NVIDIA,Aalto University, in United States of America, in June 2021. The organisation is categorised as industry,Academia.

It works in Image generation, and is recorded as doing image generation.

Because its weights were released, nothing about running it depends on a provider staying available — it is yours once downloaded.

Understanding the speeds

Across every card that can run it, the middle of the range is about 60,216.2 tokens per second, and 818 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.

Its internal architecture is not on file, so memory is approximated from the parameter count and marked accordingly. Expect the real figure to differ, more so at long context.

How it was trained

The training run consumed about 2.4 × 10²¹ FLOP, on NVIDIA V100. That figure describes the cost of creating it and has no bearing on how quickly it generates text.

Around 50,000,000 tokens went into training it.

The reason it appears in this catalogue at all is historical significance,Highly cited.

Step by step

How to choose a GPU for StyleGAN3-R

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

01

Check what it needs before anything else

The table lists every card that can hold StyleGAN3-R — around 0.7 GB at Q8_0. That figure, not the card's headline performance, is what decides whether it runs.
02

Set the context length you will work at

The conversation occupies memory too, and grows as it goes. Set the slider to the length you expect: at long context StyleGAN3-R can slip off a card that handles short questions easily.
03

Choose how far you will compress it

The quantisation column varies by card, because a bigger card holds a more accurate copy of StyleGAN3-R — Q8_0 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 StyleGAN3-R is effectively an ordering by memory bandwidth, which is why the B200 tops it at 2,144,453 tok/s.
05

Check the fit verdict before buying

Tight means StyleGAN3-R loads and works, with no room to raise the context later. Comfortable means you can. The difference matters more than a few tokens per second.
06

Open the card you have settled on

Each card page repeats this sweep for every model we hold. It answers what else the hardware is good for, beyond StyleGAN3-R.

Answers

StyleGAN3-R — common questions

Is StyleGAN3-R open source?

Its weights are published, so StyleGAN3-R 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.

How many parameters does StyleGAN3-R have?

StyleGAN3-R has 1.6M parameters. 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 StyleGAN3-R?

StyleGAN3-R was published by NVIDIA,Aalto University, based in United States of America, categorised as industry,Academia.

When was StyleGAN3-R released?

StyleGAN3-R was published in June 2021. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.

What is StyleGAN3-R used for?

StyleGAN3-R works in Image generation, and is recorded as handling image generation. These are the areas it was designed around; they describe intent rather than a hard boundary.

Where can I download StyleGAN3-R?

The weights for StyleGAN3-R are published, though we do not hold a repository link for it. This site calculates hardware requirements rather than hosting model files.

How much compute was used to train StyleGAN3-R?

Around 2.4 × 10²¹ FLOP, on NVIDIA V100. 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 StyleGAN3-R if it does not fit in my GPU?

Only by offloading, which is usually a false economy: the part in system memory drags the whole thing down. Our figures for StyleGAN3-R assume it is fully resident.

Would two GPUs run StyleGAN3-R faster?

A second card roughly doubles the memory available but not the generation rate. With 818 cards already able to run StyleGAN3-R alone, the case for pairing is weak.

Why does the quantisation differ between cards for StyleGAN3-R?

Because capacity varies, so does how hard StyleGAN3-R has to be squeezed — 1 distinct levels appear in the table above. Set a minimum quality to compare at one.

How accurate are these StyleGAN3-R speed estimates?

They are calculated from specifications rather than measured, and each carries a range — 1,286,672–3,431,124 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 StyleGAN3-R?

The smallest card in our catalogue that holds StyleGAN3-R is the Tesla C1080, with 4 GB of memory. It runs the model at Q8_0 using about 0.7 GB, and produces roughly 23,332 tokens per second. 818 cards in total can run it.

How fast is StyleGAN3-R on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 2,144,453 tokens per second; the slowest that still runs it manages considerably less. Reading speed is around ten tokens per second, and 818 of the cards that can run StyleGAN3-R clear that.

How much VRAM does StyleGAN3-R need?

About 0.7 GB at Q8_0 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 StyleGAN3-R on a 8 GB GPU?

Yes. A CMP 170HX 8 GB with 8 GB runs it at Q8_0, using about 0.7 GB and generating roughly 399,404 tokens per second — a comfortable fit.

Can I run StyleGAN3-R on a 12 GB GPU?

Yes. A GeForce RTX 3080 Ti with 12 GB runs it at Q8_0, using about 0.7 GB and generating roughly 244,575 tokens per second — a comfortable fit.

Can I run StyleGAN3-R on a 16 GB GPU?

Yes. A Tesla V100 SXM2 16 GB with 16 GB runs it at Q8_0, using about 0.7 GB and generating roughly 302,904 tokens per second — a comfortable fit.

Can I run StyleGAN3-R on a 24 GB GPU?

Yes. A GeForce RTX 5090 D V2 with 24 GB runs it at Q8_0, using about 0.7 GB and generating roughly 359,196 tokens per second — a comfortable fit.

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