ProBERTa TPS calculator

Open weights University of Illinois Urbana-Champaign (UIUC),Reed College 44M parameters September 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 · 838 tok/s

Fastest card

B200

77,005 tok/s · 180 GB

Which GPUs can run ProBERTa?

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
77,005 tok/s 46,203–123,209 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	0.7 GB	Q8_0	Comfortable
77,005 tok/s 46,203–123,209 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	0.7 GB	Q8_0	Comfortable
61,491 tok/s 36,894–98,385 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	0.7 GB	Q8_0	Comfortable
61,491 tok/s 36,894–98,385 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	0.7 GB	Q8_0	Comfortable
49,178 tok/s 29,507–78,684 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	0.7 GB	Q8_0	Comfortable
47,070 tok/s 28,242–75,311 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	0.7 GB	Q8_0	Comfortable
47,070 tok/s 28,242–75,311 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	0.7 GB	Q8_0	Comfortable
45,048 tok/s 27,029–72,077 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	0.7 GB	Q8_0	Comfortable
39,980 tok/s 23,988–63,968 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	0.7 GB	Q8_0	Comfortable
39,980 tok/s 23,988–63,968 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	0.7 GB	Q8_0	Comfortable
39,980 tok/s 23,988–63,968 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	0.7 GB	Q8_0	Comfortable
37,925 tok/s 22,755–60,680 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
32,342 tok/s 19,405–51,748 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
32,342 tok/s 19,405–51,748 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	0.7 GB	Q8_0	Comfortable
32,342 tok/s 19,405–51,748 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
32,342 tok/s 19,405–51,748 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
32,342 tok/s 19,405–51,748 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
24,626 tok/s 14,776–39,402 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	0.7 GB	Q8_0	Comfortable
24,626 tok/s 14,776–39,402 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	0.7 GB	Q8_0	Comfortable
20,522 tok/s 12,313–32,835 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	0.7 GB	Q8_0	Comfortable
20,084 tok/s 12,050–32,134 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	0.7 GB	Q8_0	Comfortable
19,636 tok/s 11,782–31,418 · low confidence	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	0.7 GB	Q8_0	Comfortable
19,636 tok/s 11,782–31,418 · low confidence	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	0.7 GB	Q8_0	Comfortable
19,636 tok/s 11,782–31,418 · low confidence	A800 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Aug 2022	0.7 GB	Q8_0	Comfortable
19,636 tok/s 11,782–31,418 · 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: University of Illinois Urbana-Champaign (UIUC),Reed College
Organisation type: Academia,Academia
Country: United States of America
Published: 1 September 2020
Authors: Ananthan Nambiar, Maeve Heflin, Simon Liu, Sergei Maslov, Mark Hopkins, Anna Ritz

What it does

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

Domain: Biology
Task: Proteins, Protein representation learning, Protein classification, Protein interaction prediction
Approach: Self-supervised learning
Numerical format: FP16

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: 44M
Training data: 58,320,000 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: 9.7 × 10¹⁸ 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: 4
Wall-clock time: 18 hours
Power draw: 2.4 kW
Compute cost: $26

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: SOTA improvement
Record confidence: Confident
Citations: 97

Sources

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

Reference: Transforming the Language of Life: Transformer Neural Networks for Protein Prediction Tasks
Last updated: 28 November 2025

The extremes

The ten fastest GPUs that run ProBERTa

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 ProBERTa

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

77,005 tok/s

ProBERTa is small enough at 44M 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 838 tokens per second.

A B200 is the fastest we calculate for it: about 77,005 tokens per second, from 8,000 GB/s of memory bandwidth.

About this model

ProBERTa was published by University of Illinois Urbana-Champaign (UIUC),Reed College, in United States of America, in September 2020. academia,Academia is the category the publisher falls under.

It works in Biology, and is recorded as doing proteins, Protein representation learning, Protein classification, Protein interaction prediction.

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

How fast it runs, and why

The median result is around 2,162.3 tokens per second; 818 cards produce text faster than most people read it.

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.

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

Producing it required around 9.7 × 10¹⁸ FLOP of arithmetic, on NVIDIA V100, which is a statement about the training budget rather than about inference.

The training set ran to roughly 58,320,000 tokens.

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

Step by step

How to choose a GPU for ProBERTa

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 ProBERTa — around 0.7 GB at Q8_0. That figure, not the card's headline performance, is what decides whether it runs.
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 ProBERTa.
03

Choose how far you will compress it

Each card runs the least-compressed copy it can hold — Q8_0 on the smallest card that fits. Setting a floor drops the cards that only manage ProBERTa by squeezing it further than you would want.
04

Rank by throughput rather than spec sheet

Ranking by tokens per second for ProBERTa follows memory bandwidth, not core counts, which is why the B200 tops it at 77,005 tok/s.
05

Read the fit column last

Tight means ProBERTa 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

Every card name links to its own page, which runs the same calculation across the whole model catalogue. Worth a look before buying for ProBERTa alone — a card is usually bought for more than one model.

Answers

ProBERTa — common questions

Can I run ProBERTa 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 8,782 tokens per second — a comfortable fit.

Can I run ProBERTa 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 10,877 tokens per second — a comfortable fit.

Can I run ProBERTa 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 12,898 tokens per second — a comfortable fit.

Is ProBERTa open source?

Its weights are published, so ProBERTa 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 ProBERTa have?

ProBERTa has 44M parameters. "In total, our model has approximately 44M trainable 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 ProBERTa?

ProBERTa was published by University of Illinois Urbana-Champaign (UIUC),Reed College, based in United States of America, categorised as academia,Academia.

When was ProBERTa released?

ProBERTa was published in September 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 ProBERTa used for?

ProBERTa works in Biology, and is recorded as handling proteins, Protein representation learning, Protein classification, Protein interaction prediction. Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.

Where can I download ProBERTa?

The weights for ProBERTa 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 ProBERTa?

Around 9.7 × 10¹⁸ 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 ProBERTa if it does not fit in my GPU?

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

Would two GPUs run ProBERTa faster?

Capacity adds across cards; throughput does not. Since 818 of the cards we track already hold ProBERTa on their own, a second card is rarely the answer here.

Why does the quantisation differ between cards for ProBERTa?

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

How accurate are these ProBERTa speed estimates?

These are estimates with real error bars. The fastest result here, 46,203–123,209 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 ProBERTa?

The smallest card in our catalogue that holds ProBERTa is the Tesla C1080, with 4 GB of memory. It runs the model at Q8_0 using about 0.7 GB, and produces roughly 838 tokens per second. 818 cards in total can run it.

How fast is ProBERTa on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 77,005 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 ProBERTa clear that.

How much VRAM does ProBERTa 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 ProBERTa 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 14,342 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