ProtBERT-BFD TPS calculator

Open weights Technical University of Munich,NVIDIA,Seoul National University,Google,Oak Ridge National Laboratory,Med AI Technology 420M parameters May 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 · 87.8 tok/s

Fastest card

B200

8,067 tok/s · 180 GB

Which GPUs can run ProtBERT-BFD?

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
8,067 tok/s 4,840–12,908 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	1.1 GB	Q8_0	Comfortable
8,067 tok/s 4,840–12,908 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	1.1 GB	Q8_0	Comfortable
6,442 tok/s 3,865–10,307 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	1.1 GB	Q8_0	Comfortable
6,442 tok/s 3,865–10,307 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	1.1 GB	Q8_0	Comfortable
5,152 tok/s 3,091–8,243 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	1.1 GB	Q8_0	Comfortable
4,931 tok/s 2,959–7,890 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	1.1 GB	Q8_0	Comfortable
4,931 tok/s 2,959–7,890 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	1.1 GB	Q8_0	Comfortable
4,719 tok/s 2,832–7,551 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	1.1 GB	Q8_0	Comfortable
4,188 tok/s 2,513–6,701 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	1.1 GB	Q8_0	Comfortable
4,188 tok/s 2,513–6,701 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	1.1 GB	Q8_0	Comfortable
4,188 tok/s 2,513–6,701 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	1.1 GB	Q8_0	Comfortable
3,973 tok/s 2,384–6,357 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	1.1 GB	Q8_0	Comfortable
3,388 tok/s 2,033–5,421 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	1.1 GB	Q8_0	Comfortable
3,388 tok/s 2,033–5,421 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	1.1 GB	Q8_0	Comfortable
3,388 tok/s 2,033–5,421 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	1.1 GB	Q8_0	Comfortable
3,388 tok/s 2,033–5,421 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	1.1 GB	Q8_0	Comfortable
3,388 tok/s 2,033–5,421 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	1.1 GB	Q8_0	Comfortable
2,580 tok/s 1,548–4,128 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	1.1 GB	Q8_0	Comfortable
2,580 tok/s 1,548–4,128 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	1.1 GB	Q8_0	Comfortable
2,150 tok/s 1,290–3,440 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	1.1 GB	Q8_0	Comfortable
2,104 tok/s 1,262–3,366 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	1.1 GB	Q8_0	Comfortable
2,057 tok/s 1,234–3,291 · low confidence	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	1.1 GB	Q8_0	Comfortable
2,057 tok/s 1,234–3,291 · low confidence	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	1.1 GB	Q8_0	Comfortable
2,057 tok/s 1,234–3,291 · low confidence	A800 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Aug 2022	1.1 GB	Q8_0	Comfortable
2,057 tok/s 1,234–3,291 · low confidence	H100 CNX NVIDIA	80 GB	2,040 GB/s	Mar 2023	1.1 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: Technical University of Munich,NVIDIA,Seoul National University,Google,Oak Ridge National Laboratory,Med AI Technology
Organisation type: Academia,Industry,Academia,Industry,Government
Country: Germany, United States of America, Korea (Republic of), China
Published: 4 May 2021
Authors: Ahmed Elnaggar, Michael Heinzinger, Christian Dallago, Ghalia Rehawi, Yu Wang, Llion Jones, Tom Gibbs, Tamas Feher, Christoph Angerer, Martin Steinegger, Debsindhu Bhowmik, Burkhard Rost

What it does

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

Domain: Biology
Task: Proteins, Protein or nucleotide language model (pLM/nLM)
Approach: Self-supervised learning
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: 420M
Training data: 58,950,000,000 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: 3.9 × 10²² FLOP
How it was established: Operation counting

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: 1,024
Power draw: 933.1 kW
Compute cost: $45,351

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

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: SOTA improvement
Record confidence: Confident

Sources

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

Reference: ProtTrans:Towards Cracking the Language of Life's Code Through Self-Supervised Learning
Last updated: 28 November 2025

The extremes

The ten fastest GPUs that run ProtBERT-BFD

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 ProtBERT-BFD

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

1.1 GB

Fastest

8,067 tok/s

ProtBERT-BFD is small enough at 420M 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 87.8 tokens per second.

A B200 is the fastest we calculate for it: about 8,067 tokens per second, from 8,000 GB/s of memory bandwidth.

About this model

ProtBERT-BFD was published by Technical University of Munich,NVIDIA,Seoul National University,Google,Oak Ridge National Laboratory,Med AI Technology, in Germany, in May 2021. It comes out of academia,Industry,Academia,Industry,Government.

It works in Biology, and is recorded as doing proteins, Protein or nucleotide language model (pLM/nLM).

The weights being open is what puts this page in the calculator rather than only in the catalogue: it is a model you can actually hold.

How fast it runs, and why

Half the cards that hold it manage more than 226.5 tokens per second, and 817 exceed reading speed outright.

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.

What went into building it

Training it took roughly 3.9 × 10²² FLOP of computation, on Google TPU v3 — a measure of what producing the model cost, not of how fast it answers.

Around 58,950,000,000 tokens went into training it.

The reason it appears in this catalogue at all is sOTA improvement.

Step by step

How to choose a GPU for ProtBERT-BFD

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

01

Read the memory figure first

Look at what ProtBERT-BFD actually needs — around 1.1 GB at Q8_0. No amount of processing power compensates for a card that cannot hold it.
02

Decide how long your conversations run

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

Set a quality floor

Compression is what makes ProtBERT-BFD 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

Compare tokens per second, not specifications

Ranking by tokens per second for ProtBERT-BFD follows memory bandwidth, not core counts, which is why the B200 tops it at 8,067 tok/s.
05

Check the fit verdict before buying

Tight means ProtBERT-BFD 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

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

Answers

ProtBERT-BFD — common questions

Can I run ProtBERT-BFD 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 ProtBERT-BFD assume it is fully resident.

Would two GPUs run ProtBERT-BFD faster?

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

Why does the quantisation differ between cards for ProtBERT-BFD?

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

How accurate are these ProtBERT-BFD speed estimates?

These are estimates with real error bars. The fastest result here, 4,840–12,908 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 ProtBERT-BFD?

The smallest card in our catalogue that holds ProtBERT-BFD is the Tesla C1080, with 4 GB of memory. It runs the model at Q8_0 using about 1.1 GB, and produces roughly 87.8 tokens per second. 818 cards in total can run it.

How fast is ProtBERT-BFD on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 8,067 tokens per second; the slowest that still runs it manages considerably less. Reading speed is around ten tokens per second, and 817 of the cards that can run ProtBERT-BFD clear that.

How much VRAM does ProtBERT-BFD need?

About 1.1 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 ProtBERT-BFD on a 8 GB GPU?

Yes. A CMP 170HX 8 GB with 8 GB runs it at Q8_0, using about 1.1 GB and generating roughly 1,503 tokens per second — a comfortable fit.

Can I run ProtBERT-BFD on a 12 GB GPU?

Yes. A GeForce RTX 3080 Ti with 12 GB runs it at Q8_0, using about 1.1 GB and generating roughly 920 tokens per second — a comfortable fit.

Can I run ProtBERT-BFD on a 16 GB GPU?

Yes. A Tesla V100 SXM2 16 GB with 16 GB runs it at Q8_0, using about 1.1 GB and generating roughly 1,140 tokens per second — a comfortable fit.

Can I run ProtBERT-BFD on a 24 GB GPU?

Yes. A GeForce RTX 5090 D V2 with 24 GB runs it at Q8_0, using about 1.1 GB and generating roughly 1,351 tokens per second — a comfortable fit.

Is ProtBERT-BFD open source?

Its weights are published, so ProtBERT-BFD 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 ProtBERT-BFD have?

ProtBERT-BFD has 420M parameters. Table 2. 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 ProtBERT-BFD?

ProtBERT-BFD was published by Technical University of Munich,NVIDIA,Seoul National University,Google,Oak Ridge National Laboratory,Med AI Technology, based in Germany, categorised as academia,Industry,Academia,Industry,Government.

When was ProtBERT-BFD released?

ProtBERT-BFD was published in May 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 ProtBERT-BFD used for?

ProtBERT-BFD works in Biology, and is recorded as handling proteins, Protein or nucleotide language model (pLM/nLM). Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.

Where can I download ProtBERT-BFD?

The weights for ProtBERT-BFD 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 ProtBERT-BFD?

Around 3.9 × 10²² 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.

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