TinyBert TPS calculator

Open weights Huazhong University of Science and Technology,Huawei Noah's Ark Lab,Huawei 67M parameters October 2020

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 · 550 tok/s

Fastest card

B200

50,571 tok/s · 180 GB

Which GPUs can run TinyBert?

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
50,571 tok/s 30,342–80,913 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	0.8 GB	Q8_0	Comfortable
50,571 tok/s 30,342–80,913 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	0.8 GB	Q8_0	Comfortable
40,382 tok/s 24,229–64,611 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	0.8 GB	Q8_0	Comfortable
40,382 tok/s 24,229–64,611 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	0.8 GB	Q8_0	Comfortable
32,296 tok/s 19,377–51,673 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	0.8 GB	Q8_0	Comfortable
30,911 tok/s 18,547–49,458 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	0.8 GB	Q8_0	Comfortable
30,911 tok/s 18,547–49,458 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	0.8 GB	Q8_0	Comfortable
29,584 tok/s 17,750–47,334 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	0.8 GB	Q8_0	Comfortable
26,256 tok/s 15,753–42,009 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	0.8 GB	Q8_0	Comfortable
26,256 tok/s 15,753–42,009 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	0.8 GB	Q8_0	Comfortable
26,256 tok/s 15,753–42,009 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	0.8 GB	Q8_0	Comfortable
24,906 tok/s 14,944–39,850 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	0.8 GB	Q8_0	Comfortable
21,240 tok/s 12,744–33,983 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	0.8 GB	Q8_0	Comfortable
21,240 tok/s 12,744–33,983 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	0.8 GB	Q8_0	Comfortable
21,240 tok/s 12,744–33,983 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	0.8 GB	Q8_0	Comfortable
21,240 tok/s 12,744–33,983 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	0.8 GB	Q8_0	Comfortable
21,240 tok/s 12,744–33,983 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	0.8 GB	Q8_0	Comfortable
16,173 tok/s 9,704–25,876 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	0.8 GB	Q8_0	Comfortable
16,173 tok/s 9,704–25,876 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	0.8 GB	Q8_0	Comfortable
13,477 tok/s 8,086–21,563 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	0.8 GB	Q8_0	Comfortable
13,189 tok/s 7,914–21,103 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	0.8 GB	Q8_0	Comfortable
12,896 tok/s 7,737–20,633 · low confidence	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	0.8 GB	Q8_0	Comfortable
12,896 tok/s 7,737–20,633 · low confidence	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	0.8 GB	Q8_0	Comfortable
12,896 tok/s 7,737–20,633 · low confidence	A800 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Aug 2022	0.8 GB	Q8_0	Comfortable
12,896 tok/s 7,737–20,633 · 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: Huazhong University of Science and Technology,Huawei Noah's Ark Lab,Huawei
Organisation type: Academia,Industry,Industry
Country: China
Published: 16 October 2020
Authors: Xiaoqi Jiao, Yichun Yin, Lifeng Shang, Xin Jiang, Xiao Chen, Linlin Li, Fang Wang, Qun Liu

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

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: 67M
Training data: 3,300,000,000 tokens
Epochs: 3

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: 4 × 10¹⁸ 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: 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: Confident

Sources

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

Reference: TinyBERT: Distilling BERT for Natural Language Understanding
Last updated: 28 November 2025

The extremes

The ten fastest GPUs that run TinyBert

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 TinyBert

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

Tesla C1080

Memory needed

0.8 GB

Fastest

50,571 tok/s

TinyBert is small enough at 67M parameters that hardware is rarely the obstacle — 818 of the cards we track can run it, including cards several years old.

The entry point is the Tesla C1080: 4 GB of memory, Q8_0 compression, roughly 550 tokens per second.

A B200 is the fastest we calculate for it: about 50,571 tokens per second, from 8,000 GB/s of memory bandwidth.

What this model is

TinyBert was published by Huazhong University of Science and Technology,Huawei Noah's Ark Lab,Huawei, in China, in October 2020. It comes out of academia,Industry,Industry.

It works in Language, and is recorded as doing language modeling/generation, Question answering.

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.

What decides the speed

Across every card that can run it, the middle of the range is about 1,420.0 tokens per second, and 818 of them clear the ten tokens per second that roughly matches reading speed.

It is a dense model, so every parameter is read for every token produced. That makes speed track memory bandwidth almost exactly — a card with twice the bandwidth generates roughly twice as fast.

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.

What went into building it

Training it took roughly 4 × 10¹⁸ FLOP of computation — a measure of what producing the model cost, not of how fast it answers.

The training set ran to roughly 3,300,000,000 tokens.

Step by step

How to choose a GPU for TinyBert

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 TinyBert 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 TinyBert 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 TinyBert — Q8_0 on the smallest card that fits. Set a floor to hold the comparison at one level.
04

Compare tokens per second, not specifications

Sort by speed to see how cards rank for TinyBert. It will not match a gaming ordering — generation is bound by memory bandwidth, which is why the B200 tops it at 50,571 tok/s.
05

Look at the headroom, not just the fit

A tight fit runs TinyBert 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

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

Answers

TinyBert — common questions

Can I run TinyBert 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 7,143 tokens per second — a comfortable fit.

Can I run TinyBert 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 8,471 tokens per second — a comfortable fit.

Is TinyBert open source?

Its weights are published, so TinyBert 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 TinyBert have?

TinyBert has 67M parameters. 67M. 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 TinyBert?

TinyBert was published by Huazhong University of Science and Technology,Huawei Noah's Ark Lab,Huawei, based in China, categorised as academia,Industry,Industry.

When was TinyBert released?

TinyBert was published in October 2020. 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 TinyBert used for?

TinyBert works in Language, and is recorded as handling language modeling/generation, Question answering. These are the areas it was designed around; they describe intent rather than a hard boundary.

Where can I download TinyBert?

The weights for TinyBert 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 TinyBert?

Around 4 × 10¹⁸ 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 TinyBert 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 TinyBert is rarely worth using. Every figure here assumes the whole model is on the card.

Would two GPUs run TinyBert faster?

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

Why does the quantisation differ between cards for TinyBert?

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

How accurate are these TinyBert speed estimates?

They are calculated from specifications rather than measured, and each carries a range — 30,342–80,913 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 TinyBert?

The smallest card in our catalogue that holds TinyBert is the Tesla C1080, with 4 GB of memory. It runs the model at Q8_0 using about 0.8 GB, and produces roughly 550 tokens per second. 818 cards in total can run it.

How fast is TinyBert on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 50,571 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 TinyBert clear that.

How much VRAM does TinyBert 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 TinyBert 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 9,419 tokens per second — a comfortable fit.

Can I run TinyBert 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 5,768 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