DeiT-B TPS calculator

Open weights Meta AI,Sorbonne University 86M parameters January 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 · 429 tok/s

Fastest card

B200

39,398 tok/s · 180 GB

Which GPUs can run DeiT-B?

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
39,398 tok/s 23,639–63,037 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	0.8 GB	Q8_0	Comfortable
39,398 tok/s 23,639–63,037 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	0.8 GB	Q8_0	Comfortable
31,460 tok/s 18,876–50,337 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	0.8 GB	Q8_0	Comfortable
31,460 tok/s 18,876–50,337 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	0.8 GB	Q8_0	Comfortable
25,161 tok/s 15,096–40,257 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	0.8 GB	Q8_0	Comfortable
24,082 tok/s 14,449–38,531 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	0.8 GB	Q8_0	Comfortable
24,082 tok/s 14,449–38,531 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	0.8 GB	Q8_0	Comfortable
23,048 tok/s 13,829–36,877 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	0.8 GB	Q8_0	Comfortable
20,455 tok/s 12,273–32,728 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	0.8 GB	Q8_0	Comfortable
20,455 tok/s 12,273–32,728 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	0.8 GB	Q8_0	Comfortable
20,455 tok/s 12,273–32,728 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	0.8 GB	Q8_0	Comfortable
19,404 tok/s 11,642–31,046 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	0.8 GB	Q8_0	Comfortable
16,547 tok/s 9,928–26,476 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	0.8 GB	Q8_0	Comfortable
16,547 tok/s 9,928–26,476 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	0.8 GB	Q8_0	Comfortable
16,547 tok/s 9,928–26,476 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	0.8 GB	Q8_0	Comfortable
16,547 tok/s 9,928–26,476 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	0.8 GB	Q8_0	Comfortable
16,547 tok/s 9,928–26,476 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	0.8 GB	Q8_0	Comfortable
12,600 tok/s 7,560–20,159 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	0.8 GB	Q8_0	Comfortable
12,600 tok/s 7,560–20,159 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	0.8 GB	Q8_0	Comfortable
10,500 tok/s 6,300–16,799 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	0.8 GB	Q8_0	Comfortable
10,276 tok/s 6,165–16,441 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	0.8 GB	Q8_0	Comfortable
10,047 tok/s 6,028–16,074 · low confidence	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	0.8 GB	Q8_0	Comfortable
10,047 tok/s 6,028–16,074 · low confidence	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	0.8 GB	Q8_0	Comfortable
10,047 tok/s 6,028–16,074 · low confidence	A800 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Aug 2022	0.8 GB	Q8_0	Comfortable
10,047 tok/s 6,028–16,074 · low confidence	H100 CNX NVIDIA	80 GB	2,040 GB/s	Mar 2023	0.8 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: Meta AI,Sorbonne University
Organisation type: Industry,Academia
Country: United States of America, France
Published: 15 January 2021
Authors: Hugo Touvron, Matthieu Cord, Matthijs Douze, Francisco Massa, Alexandre Sablayrolles, Hervé Jégou

What it does

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

Domain: Vision
Task: Image classification

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: 86M
Training data: 3,840,000 tokens
Epochs: 300

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: 7.9 × 10¹⁹ 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
Wall-clock time: 53 hours

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.

Why it is tracked: Highly cited
Record confidence: Confident
Citations: 9,077

Sources

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

Reference: Training data-efficient image transformers & distillation through attention
Last updated: 25 May 2026

The extremes

The ten fastest GPUs that run DeiT-B

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 DeiT-B

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

0.8 GB

Fastest

39,398 tok/s

DeiT-B is small enough at 86M 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 429 tokens per second.

At the other end, a B200 generates roughly 39,398 tokens per second on it, on the strength of 8,000 GB/s of memory bandwidth.

Background

DeiT-B was published by Meta AI,Sorbonne University, in United States of America, in January 2021. It comes out of industry,Academia.

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

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

Reading the throughput figures

Across every card that can run it, the middle of the range is about 1,106.3 tokens per second, and 818 of them clear the ten tokens per second that roughly matches reading speed.

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

Memory here is estimated from size rather than computed from the architecture, which is not recorded for this model — the numbers are indicative rather than exact.

What went into building it

Producing it required around 7.9 × 10¹⁹ FLOP of arithmetic, on NVIDIA V100, which is a statement about the training budget rather than about inference.

The training set ran to roughly 3,840,000 tokens.

The reason it appears in this catalogue at all is highly cited.

Step by step

How to choose a GPU for DeiT-B

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

Look at what DeiT-B actually needs — around 0.8 GB at Q8_0. No amount of processing power compensates for a card that cannot hold it.
02

Decide how long your conversations run

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

Set a quality floor

The quantisation column varies by card, because a bigger card holds a more accurate copy of DeiT-B — Q8_0 on the smallest card that fits. Set a floor to hold the comparison at one level.
04

Sort by speed

The speed ordering for DeiT-B is effectively an ordering by memory bandwidth, which is why the B200 tops it at 39,398 tok/s.
05

Look at the headroom, not just the fit

Tight means DeiT-B 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

Check the card from the other side

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

Answers

DeiT-B — common questions

How fast is DeiT-B on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 39,398 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 DeiT-B clear that.

How much VRAM does DeiT-B need?

About 0.8 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 DeiT-B on a 8 GB GPU?

Yes. A CMP 170HX 8 GB with 8 GB runs it at Q8_0, using about 0.8 GB and generating roughly 7,338 tokens per second — a comfortable fit.

Can I run DeiT-B on a 12 GB GPU?

Yes. A GeForce RTX 3080 Ti with 12 GB runs it at Q8_0, using about 0.8 GB and generating roughly 4,493 tokens per second — a comfortable fit.

Can I run DeiT-B on a 16 GB GPU?

Yes. A Tesla V100 SXM2 16 GB with 16 GB runs it at Q8_0, using about 0.8 GB and generating roughly 5,565 tokens per second — a comfortable fit.

Can I run DeiT-B on a 24 GB GPU?

Yes. A GeForce RTX 5090 D V2 with 24 GB runs it at Q8_0, using about 0.8 GB and generating roughly 6,599 tokens per second — a comfortable fit.

Is DeiT-B open source?

Its weights are published, so DeiT-B 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 DeiT-B have?

DeiT-B has 86M parameters. (DeiT-B). 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 DeiT-B?

DeiT-B was published by Meta AI,Sorbonne University, based in United States of America, categorised as industry,Academia.

When was DeiT-B released?

DeiT-B was published in January 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 DeiT-B used for?

DeiT-B works in Vision, and is recorded as handling image classification. A model can carry several of each, so these are the areas it was built for rather than a limit on what it will attempt.

Where can I download DeiT-B?

The weights for DeiT-B 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 DeiT-B?

Around 7.9 × 10¹⁹ 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 DeiT-B 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 DeiT-B assume it is fully resident.

Would two GPUs run DeiT-B faster?

Two cards buy memory rather than speed. That matters for DeiT-B only if one card cannot hold it — 818 can, so a second adds little.

Why does the quantisation differ between cards for DeiT-B?

A larger card holds a more accurate copy. Across the cards that run DeiT-B, 1 compression levels are used; the floor control above pins it to one.

How accurate are these DeiT-B speed estimates?

They are calculated from specifications rather than measured, and each carries a range — 23,639–63,037 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 DeiT-B?

The smallest card in our catalogue that holds DeiT-B is the Tesla C1080, with 4 GB of memory. It runs the model at Q8_0 using about 0.8 GB, and produces roughly 429 tokens per second. 818 cards in total can run it.

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