MnasNet-A3 TPS calculator

Open weights Google 5.2M parameters May 2019

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 · 7,089 tok/s

Fastest card

B200

651,584 tok/s · 180 GB

Which GPUs can run MnasNet-A3?

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
651,584 tok/s 390,950–1,042,534 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	0.7 GB	Q8_0	Comfortable
651,584 tok/s 390,950–1,042,534 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	0.7 GB	Q8_0	Comfortable
520,306 tok/s 312,184–832,489 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	0.7 GB	Q8_0	Comfortable
520,306 tok/s 312,184–832,489 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	0.7 GB	Q8_0	Comfortable
416,118 tok/s 249,671–665,788 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	0.7 GB	Q8_0	Comfortable
398,281 tok/s 238,968–637,249 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	0.7 GB	Q8_0	Comfortable
398,281 tok/s 238,968–637,249 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	0.7 GB	Q8_0	Comfortable
381,176 tok/s 228,706–609,882 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	0.7 GB	Q8_0	Comfortable
338,294 tok/s 202,976–541,271 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	0.7 GB	Q8_0	Comfortable
338,294 tok/s 202,976–541,271 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	0.7 GB	Q8_0	Comfortable
338,294 tok/s 202,976–541,271 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	0.7 GB	Q8_0	Comfortable
320,905 tok/s 192,543–513,448 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
273,665 tok/s 164,199–437,864 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
273,665 tok/s 164,199–437,864 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	0.7 GB	Q8_0	Comfortable
273,665 tok/s 164,199–437,864 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
273,665 tok/s 164,199–437,864 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
273,665 tok/s 164,199–437,864 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
208,376 tok/s 125,026–333,402 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	0.7 GB	Q8_0	Comfortable
208,376 tok/s 125,026–333,402 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	0.7 GB	Q8_0	Comfortable
173,647 tok/s 104,188–277,835 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	0.7 GB	Q8_0	Comfortable
169,941 tok/s 101,965–271,906 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	0.7 GB	Q8_0	Comfortable
166,154 tok/s 99,692–265,846 · low confidence	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	0.7 GB	Q8_0	Comfortable
166,154 tok/s 99,692–265,846 · low confidence	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	0.7 GB	Q8_0	Comfortable
166,154 tok/s 99,692–265,846 · low confidence	A800 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Aug 2022	0.7 GB	Q8_0	Comfortable
166,154 tok/s 99,692–265,846 · 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: Google
Organisation type: Industry
Country: United States of America
Published: 29 May 2019
Authors: Mingxing Tan, Bo Chen, Ruoming Pang, Vijay Vasudevan, Mark Sandler, Andrew Howard, Quoc V. Le

What it does

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

Domain: Vision
Task: Image classification, Object detection
Numerical format: FP32

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: 5.2M
Training data: 1,230,000 tokens

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: 1.5 × 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: Google TPU v3
Chips used: 256
Wall-clock time: 108 hours
Power draw: 237.0 kW
Compute cost: $9,552

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: Open source

How it is classified

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

Record confidence: Speculative
Citations: 3,396

Sources

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

Reference: MnasNet: Platform-Aware Neural Architecture Search for Mobile
Last updated: 25 May 2026

The extremes

The ten fastest GPUs that run MnasNet-A3

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 MnasNet-A3

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

The hardware side

Minimum card

Tesla C1080

Memory needed

0.7 GB

Fastest

651,584 tok/s

MnasNet-A3 is small enough at 5.2M parameters that hardware is rarely the obstacle — 818 of the cards we track can run it, including cards several years old.

At the low end, a Tesla C1080 handles it — 4 GB, at Q8_0, for about 7,089 tokens per second.

Top of the range is the B200, at roughly 651,584 tokens per second thanks to 8,000 GB/s of bandwidth.

Background

MnasNet-A3 was published by Google, in United States of America, in May 2019. industry is the category the publisher falls under.

It works in Vision, and is recorded as doing image classification, Object detection.

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.

Reading the throughput figures

Half the cards that hold it manage more than 18,296.5 tokens per second, and 818 exceed reading speed outright.

Every weight participates in every token here, so bandwidth is the whole story: the ranking below is effectively a ranking of memory throughput.

Without the attention layout on record, the memory column is an approximation. It is close enough to choose hardware by, and least reliable at long context.

Training and provenance

Training it took roughly 1.5 × 10²¹ FLOP of computation, on Google TPU v3 — a measure of what producing the model cost, not of how fast it answers.

It was trained on about 1,230,000 tokens of text.

Step by step

How to choose a GPU for MnasNet-A3

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 MnasNet-A3 — 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

Longer conversations cost memory on top of what the weights need. Move the slider to your real working length before trusting any row for MnasNet-A3.
03

Decide how much compression you will accept

Compression is what makes MnasNet-A3 fit smaller cards, at some cost in accuracy — Q8_0 on the smallest card that fits. A minimum quality removes the ones that go too far.
04

Rank by throughput rather than spec sheet

The speed ordering for MnasNet-A3 is effectively an ordering by memory bandwidth, which is why the B200 tops it at 651,584 tok/s.
05

Look at the headroom, not just the fit

A tight fit runs MnasNet-A3 but leaves nothing spare for a longer conversation; comfortable has headroom. If you expect to grow the context, buy for comfortable.
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 MnasNet-A3 is settled.

Answers

MnasNet-A3 — common questions

Why does the quantisation differ between cards for MnasNet-A3?

Because capacity varies, so does how hard MnasNet-A3 has to be squeezed — 1 distinct levels appear in the table above. Set a minimum quality to compare at one.

How accurate are these MnasNet-A3 speed estimates?

These are estimates with real error bars. The fastest result here, 390,950–1,042,534 tok/s on the B200, could reasonably land anywhere in its published range depending on which runtime you use.

What GPU do I need to run MnasNet-A3?

The smallest card in our catalogue that holds MnasNet-A3 is the Tesla C1080, with 4 GB of memory. It runs the model at Q8_0 using about 0.7 GB, and produces roughly 7,089 tokens per second. 818 cards in total can run it.

How fast is MnasNet-A3 on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 651,584 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 MnasNet-A3 clear that.

How much VRAM does MnasNet-A3 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 MnasNet-A3 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 121,357 tokens per second — a comfortable fit.

Can I run MnasNet-A3 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 74,313 tokens per second — a comfortable fit.

Can I run MnasNet-A3 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 92,036 tokens per second — a comfortable fit.

Can I run MnasNet-A3 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 109,140 tokens per second — a comfortable fit.

Is MnasNet-A3 open source?

Its weights are published, so MnasNet-A3 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 MnasNet-A3 have?

MnasNet-A3 has 5.2M parameters. From https://arxiv.org/pdf/1807.11626.pdf. 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 MnasNet-A3?

MnasNet-A3 was published by Google, based in United States of America, categorised as industry.

When was MnasNet-A3 released?

MnasNet-A3 was published in May 2019. 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 MnasNet-A3 used for?

MnasNet-A3 works in Vision, and is recorded as handling image classification, Object detection. Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.

Where can I download MnasNet-A3?

The weights for MnasNet-A3 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 MnasNet-A3?

Around 1.5 × 10²¹ FLOP, on Google TPU v3. 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 MnasNet-A3 if it does not fit in my GPU?

It can be split between the card and system memory, but MnasNet-A3 generates painfully slowly that way. Nothing on this page assumes offloading.

Would two GPUs run MnasNet-A3 faster?

A second card roughly doubles the memory available but not the generation rate. With 818 cards already able to run MnasNet-A3 alone, the case for pairing is weak.

Source

Original publication

Record last updated 25 May 2026

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