Tranception TPS calculator

Open weights University of Oxford,Harvard Medical School,Cohere 700M parameters May 2022

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 · 52.7 tok/s

Fastest card

B200

4,840 tok/s · 180 GB

Which GPUs can run Tranception?

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
4,840 tok/s 2,904–7,745 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	1.4 GB	Q8_0	Comfortable
4,840 tok/s 2,904–7,745 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	1.4 GB	Q8_0	Comfortable
3,865 tok/s 2,319–6,184 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	1.4 GB	Q8_0	Comfortable
3,865 tok/s 2,319–6,184 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	1.4 GB	Q8_0	Comfortable
3,091 tok/s 1,855–4,946 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	1.4 GB	Q8_0	Comfortable
2,959 tok/s 1,775–4,734 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	1.4 GB	Q8_0	Comfortable
2,959 tok/s 1,775–4,734 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	1.4 GB	Q8_0	Comfortable
2,832 tok/s 1,699–4,531 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	1.4 GB	Q8_0	Comfortable
2,513 tok/s 1,508–4,021 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	1.4 GB	Q8_0	Comfortable
2,513 tok/s 1,508–4,021 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	1.4 GB	Q8_0	Comfortable
2,513 tok/s 1,508–4,021 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	1.4 GB	Q8_0	Comfortable
2,384 tok/s 1,430–3,814 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	1.4 GB	Q8_0	Comfortable
2,033 tok/s 1,220–3,253 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	1.4 GB	Q8_0	Comfortable
2,033 tok/s 1,220–3,253 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	1.4 GB	Q8_0	Comfortable
2,033 tok/s 1,220–3,253 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	1.4 GB	Q8_0	Comfortable
2,033 tok/s 1,220–3,253 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	1.4 GB	Q8_0	Comfortable
2,033 tok/s 1,220–3,253 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	1.4 GB	Q8_0	Comfortable
1,548 tok/s 929–2,477 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	1.4 GB	Q8_0	Comfortable
1,548 tok/s 929–2,477 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	1.4 GB	Q8_0	Comfortable
1,290 tok/s 774–2,064 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	1.4 GB	Q8_0	Comfortable
1,262 tok/s 757–2,020 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	1.4 GB	Q8_0	Comfortable
1,234 tok/s 741–1,975 · low confidence	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	1.4 GB	Q8_0	Comfortable
1,234 tok/s 741–1,975 · low confidence	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	1.4 GB	Q8_0	Comfortable
1,234 tok/s 741–1,975 · low confidence	A800 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Aug 2022	1.4 GB	Q8_0	Comfortable
1,234 tok/s 741–1,975 · low confidence	H100 CNX NVIDIA	80 GB	2,040 GB/s	Mar 2023	1.4 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 Oxford,Harvard Medical School,Cohere
Organisation type: Academia,Academia,Industry
Country: United Kingdom of Great Britain and Northern Ireland, United States of America, Canada
Published: 27 May 2022
Authors: Pascal Notin, Mafalda Dias, Jonathan Frazer, Javier Marchena-Hurtado, Aidan Gomez, Debora S. Marks, Yarin Gal

What it does

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

Domain: Biology
Task: Proteins, Protein pathogenicity prediction
Approach: Self-supervised learning

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: 700M
Training data: 48,230,400,000 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: 7.2 × 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: NVIDIA A100
Chips used: 64
Chip-hours: 21,504
Wall-clock time: 336 hours (14 days)
Power draw: 51.4 kW
Compute cost: $15,247

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
Hugging Face: OATML-Markslab

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: 243

Sources

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

Reference: Tranception: protein fitness prediction with autoregressive transformers and inference-time retrieval
Last updated: 25 May 2026

The extremes

The ten fastest GPUs that run Tranception

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 Tranception

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

1.4 GB

Fastest

4,840 tok/s

Tranception is small enough at 700M 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 52.7 tokens per second.

A B200 is the fastest we calculate for it: about 4,840 tokens per second, from 8,000 GB/s of memory bandwidth.

About this model

Tranception was published by University of Oxford,Harvard Medical School,Cohere, in United Kingdom of Great Britain and Northern Ireland, in May 2022. It comes out of academia,Academia,Industry.

It works in Biology, and is recorded as doing proteins, Protein pathogenicity prediction.

The weights are published, so it can be downloaded and run on your own hardware indefinitely, offline, with no account attached. It is published under the OATML-Markslab organisation on Hugging Face.

How fast it runs, and why

Across every card that can run it, the middle of the range is about 135.9 tokens per second, and 809 of them clear the ten tokens per second that roughly matches reading speed.

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.

How it was trained

The training run consumed about 7.2 × 10²¹ FLOP, on NVIDIA A100. That figure describes the cost of creating it and has no bearing on how quickly it generates text.

Around 48,230,400,000 tokens went into training it.

Its inclusion criterion is sOTA improvement.

Step by step

How to choose a GPU for Tranception

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

The table lists every card that can hold Tranception — around 1.4 GB at Q8_0. That figure, not the card's headline performance, is what decides whether it runs.
02

Match the context to your actual use

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

Set a quality floor

Compression is what makes Tranception 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

Sort by speed

The speed ordering for Tranception is effectively an ordering by memory bandwidth, which is why the B200 tops it at 4,840 tok/s.
05

Look at the headroom, not just the fit

Tight means Tranception 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

See what else that card runs

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

Answers

Tranception — common questions

Can I run Tranception 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 Tranception is rarely worth using. Every figure here assumes the whole model is on the card.

Would two GPUs run Tranception faster?

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

Why does the quantisation differ between cards for Tranception?

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

How accurate are these Tranception speed estimates?

They are calculated from specifications rather than measured, and each carries a range — 2,904–7,745 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 Tranception?

The smallest card in our catalogue that holds Tranception is the Tesla C1080, with 4 GB of memory. It runs the model at Q8_0 using about 1.4 GB, and produces roughly 52.7 tokens per second. 818 cards in total can run it.

How fast is Tranception on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 4,840 tokens per second; the slowest that still runs it manages considerably less. Reading speed is around ten tokens per second, and 809 of the cards that can run Tranception clear that.

How much VRAM does Tranception need?

About 1.4 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 Tranception on a 8 GB GPU?

Yes. A CMP 170HX 8 GB with 8 GB runs it at Q8_0, using about 1.4 GB and generating roughly 902 tokens per second — a comfortable fit.

Can I run Tranception on a 12 GB GPU?

Yes. A GeForce RTX 3080 Ti with 12 GB runs it at Q8_0, using about 1.4 GB and generating roughly 552 tokens per second — a comfortable fit.

Can I run Tranception on a 16 GB GPU?

Yes. A Tesla V100 SXM2 16 GB with 16 GB runs it at Q8_0, using about 1.4 GB and generating roughly 684 tokens per second — a comfortable fit.

Can I run Tranception on a 24 GB GPU?

Yes. A GeForce RTX 5090 D V2 with 24 GB runs it at Q8_0, using about 1.4 GB and generating roughly 811 tokens per second — a comfortable fit.

Is Tranception open source?

Its weights are published, so Tranception 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 Tranception have?

Tranception has 700M parameters. "Our largest transformer model, Tranception L, has 700M parameters and is trained on UniRef100 (Suzek et al., 2014)". 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 Tranception?

Tranception was published by University of Oxford,Harvard Medical School,Cohere, based in United Kingdom of Great Britain and Northern Ireland, categorised as academia,Academia,Industry.

When was Tranception released?

Tranception was published in May 2022. 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 Tranception used for?

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

Where can I download Tranception?

Its weights are published under the OATML-Markslab organisation on Hugging Face. We do not host model files — this site calculates what hardware is needed to run them.

How much compute was used to train Tranception?

Around 7.2 × 10²¹ FLOP, on NVIDIA A100. 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 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