AbLang (heavy sequences) TPS calculator

Open weights University of Oxford 355M parameters January 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 · 104 tok/s

Fastest card

B200

9,544 tok/s · 180 GB

Which GPUs can run AbLang (heavy sequences)?

Set the inputs, read the answer

A longer conversation needs more memory, which can push this model off smaller cards.

Hides cards that would only fit the model by compressing it below this point.

818 cards match

Calculating
Needs Quantisation Fit
9,544 tok/s

5,727–15,271 · low confidence

B200 NVIDIA 180 GB 8,000 GB/s Jan 2024 1.1 GB Q8_0 Comfortable
9,544 tok/s

5,727–15,271 · low confidence

B300 NVIDIA 288 GB 8,000 GB/s Sep 2025 1.1 GB Q8_0 Comfortable
7,621 tok/s

4,573–12,194 · low confidence

Radeon Instinct MI350X AMD 288 GB 8,190 GB/s Jan 2025 1.1 GB Q8_0 Comfortable
7,621 tok/s

4,573–12,194 · low confidence

Radeon Instinct MI355X AMD 288 GB 8,190 GB/s Jan 2025 1.1 GB Q8_0 Comfortable
6,095 tok/s

3,657–9,752 · low confidence

Radeon Instinct MI300 AMD 128 GB 6,550 GB/s Jan 2023 1.1 GB Q8_0 Comfortable
5,834 tok/s

3,500–9,334 · low confidence

H200 NVL NVIDIA 141 GB 4,890 GB/s Nov 2024 1.1 GB Q8_0 Comfortable
5,834 tok/s

3,500–9,334 · low confidence

H200 SXM 141 GB NVIDIA 141 GB 4,890 GB/s Nov 2024 1.1 GB Q8_0 Comfortable
5,583 tok/s

3,350–8,933 · low confidence

Radeon Instinct MI325X AMD 256 GB 6,000 GB/s Oct 2024 1.1 GB Q8_0 Comfortable
4,955 tok/s

2,973–7,928 · low confidence

Radeon Instinct MI300A AMD 128 GB 5,325 GB/s Dec 2023 1.1 GB Q8_0 Comfortable
4,955 tok/s

2,973–7,928 · low confidence

Radeon Instinct MI300X AMD 192 GB 5,325 GB/s Dec 2023 1.1 GB Q8_0 Comfortable
4,955 tok/s

2,973–7,928 · low confidence

Radeon Instinct MI308X AMD 192 GB 5,325 GB/s Dec 2023 1.1 GB Q8_0 Comfortable
4,701 tok/s

2,820–7,521 · low confidence

H100 NVL 94 GB NVIDIA 94 GB 3,940 GB/s Mar 2023 1.1 GB Q8_0 Comfortable
4,009 tok/s

2,405–6,414 · low confidence

H100 PCIe 96 GB NVIDIA 96 GB 3,360 GB/s Mar 2023 1.1 GB Q8_0 Comfortable
4,009 tok/s

2,405–6,414 · low confidence

H100 SXM5 80 GB NVIDIA 80 GB 3,360 GB/s Oct 2022 1.1 GB Q8_0 Comfortable
4,009 tok/s

2,405–6,414 · low confidence

H100 SXM5 94 GB NVIDIA 94 GB 3,360 GB/s Mar 2023 1.1 GB Q8_0 Comfortable
4,009 tok/s

2,405–6,414 · low confidence

H100 SXM5 96 GB NVIDIA 96 GB 3,360 GB/s Mar 2023 1.1 GB Q8_0 Comfortable
4,009 tok/s

2,405–6,414 · low confidence

H800 SXM5 NVIDIA 80 GB 3,360 GB/s Mar 2023 1.1 GB Q8_0 Comfortable
3,052 tok/s

1,831–4,884 · low confidence

Radeon Instinct MI250 AMD 128 GB 3,280 GB/s Nov 2021 1.1 GB Q8_0 Comfortable
3,052 tok/s

1,831–4,884 · low confidence

Radeon Instinct MI250X AMD 128 GB 3,280 GB/s Nov 2021 1.1 GB Q8_0 Comfortable
2,544 tok/s

1,526–4,070 · low confidence

Data Center GPU Max 1550 Intel 128 GB 3,280 GB/s Jan 2023 1.1 GB Q8_0 Comfortable
2,489 tok/s

1,494–3,983 · low confidence

Data Center GPU Max Subsystem Intel 128 GB 3,210 GB/s Jan 2023 1.1 GB Q8_0 Comfortable
2,434 tok/s

1,460–3,894 · low confidence

A100 SXM4 80 GB NVIDIA 80 GB 2,040 GB/s Nov 2020 1.1 GB Q8_0 Comfortable
2,434 tok/s

1,460–3,894 · low confidence

A100X NVIDIA 80 GB 2,040 GB/s Jun 2021 1.1 GB Q8_0 Comfortable
2,434 tok/s

1,460–3,894 · low confidence

A800 SXM4 80 GB NVIDIA 80 GB 2,040 GB/s Aug 2022 1.1 GB Q8_0 Comfortable
2,434 tok/s

1,460–3,894 · 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
University of Oxford
Organisation type
Academia
Country
United Kingdom of Great Britain and Northern Ireland
Published
22 January 2022
Authors
Tobias H Olsen, Iain H Moal, Charlotte M Deane

What it does

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

Domain
Biology
Task
Proteins, Antibody property prediction, Protein or nucleotide language model (pLM/nLM)

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
355M

"The hyperparameters were selected to be similar to those used in the RoBERTa paper (Liu et al., 2019)." Liu et al., 2019 link: https://arxiv.org/pdf/1907.11692.pdf "We begin by training RoBERTa following the BERTLARGE architecture (L = 24, H = 1024, A = 16, 355M parameters)"

Training data
2,260,275,840 tokens

Heavy Chain: 14,126,724 sequences × 160 residues = 2,260,275,840 datapoints Light Chain: 187,068 sequences × 160 residues = 29,930,880 datapoints Total: 2,260,275,840 + 29,930,880 = 2,290,206,720 datapoints (2.29B)

Epochs
20

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

BSD-3-Clause license https://github.com/oxpig/AbLang AbLang is a python package

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

"AbLang restores residues more accurately and faster than a current state-of-the-art protein language model ESM-1b, emphasizing the benefits and potential of an antibody specific language model" - SOTA improvement for a very specific task I haven't found information about SOTA claims on any standard benchmarks

Record confidence
Confident
Citations
176

Sources

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

Reference
AbLang: an antibody language model for completing antibody sequences
Last updated
11 February 2026

The extremes

The ten fastest GPUs that run AbLang (heavy sequences)

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.

  1. 01 B300 288 GB · 8,000 GB/s · Q8_0 9,544 tok/s
  2. 02 B200 180 GB · 8,000 GB/s · Q8_0 9,544 tok/s
  3. 03 Radeon Instinct MI350X 288 GB · 8,190 GB/s · Q8_0 7,621 tok/s
  4. 04 Radeon Instinct MI355X 288 GB · 8,190 GB/s · Q8_0 7,621 tok/s
  5. 05 Radeon Instinct MI300 128 GB · 6,550 GB/s · Q8_0 6,095 tok/s
  6. 06 H200 NVL 141 GB · 4,890 GB/s · Q8_0 5,834 tok/s
  7. 07 H200 SXM 141 GB 141 GB · 4,890 GB/s · Q8_0 5,834 tok/s
  8. 08 Radeon Instinct MI325X 256 GB · 6,000 GB/s · Q8_0 5,583 tok/s
  9. 09 Radeon Instinct MI300A 128 GB · 5,325 GB/s · Q8_0 4,955 tok/s
  10. 10 Radeon Instinct MI300X 192 GB · 5,325 GB/s · Q8_0 4,955 tok/s

The smallest GPUs that still run AbLang (heavy sequences)

The cheapest route in, by memory capacity. A tight fit runs the model but leaves nothing spare for a longer conversation.

  1. 01 GeForce RTX 4010 4 GB · needs 1.1 GB · Q8_0 · comfortable 115 tok/s
  2. 02 RTX A400 4 GB · needs 1.1 GB · Q8_0 · comfortable 115 tok/s
  3. 03 RTX 500 Mobile Ada Generation 4 GB · needs 1.1 GB · Q8_0 · comfortable 153 tok/s
  4. 04 GeForce RTX 3050 A Mobile 4 GB · needs 1.1 GB · Q8_0 · comfortable 229 tok/s
  5. 05 Jetson Orin Nano 4 GB 4 GB · needs 1.1 GB · Q8_0 · comfortable 40.7 tok/s
  6. 06 Radeon RX 6450M 4 GB · needs 1.1 GB · Q8_0 · comfortable 119 tok/s
  7. 07 Radeon RX 6550M 4 GB · needs 1.1 GB · Q8_0 · comfortable 134 tok/s
  8. 08 Radeon RX 6550S 4 GB · needs 1.1 GB · Q8_0 · comfortable 119 tok/s
  9. 09 Arc A310 4 GB · needs 1.1 GB · Q8_0 · comfortable 96.2 tok/s
  10. 10 Arc Pro A30M 4 GB · needs 1.1 GB · Q8_0 · comfortable 99.3 tok/s

What the numbers mean

The hardware side

Minimum card

Tesla C1080

Memory needed

1.1 GB

Fastest

9,544 tok/s

AbLang (heavy sequences) is small enough at 355M 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 104 tokens per second.

The quickest result comes from a B200 at around 9,544 tokens per second — its 8,000 GB/s of bandwidth is what buys that.

Where it came from

AbLang (heavy sequences) was published by University of Oxford, in United Kingdom of Great Britain and Northern Ireland, in January 2022. academia is the category the publisher falls under.

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

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.

Understanding the speeds

Half the cards that hold it manage more than 268.0 tokens per second, and 817 exceed reading speed outright.

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.

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

The training set ran to roughly 2,260,275,840 tokens.

Its inclusion criterion is sOTA improvement.

Step by step

How to choose a GPU for AbLang (heavy sequences)

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

  1. 01

    Start from the memory column

    Look at what AbLang (heavy sequences) actually needs — around 1.1 GB at Q8_0. No amount of processing power compensates for a card that cannot hold it.

  2. 02

    Set the context length you will work at

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

  3. 03

    Decide how much compression you will accept

    Compression is what makes AbLang (heavy sequences) 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.

  4. 04

    Sort by speed

    Ranking by tokens per second for AbLang (heavy sequences) follows memory bandwidth, not core counts, which is why the B200 tops it at 9,544 tok/s.

  5. 05

    Look at the headroom, not just the fit

    A tight fit runs AbLang (heavy sequences) but leaves nothing spare for a longer conversation; comfortable has headroom. If you expect to grow the context, buy for comfortable.

  6. 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 AbLang (heavy sequences).

Answers

AbLang (heavy sequences) — common questions

01

How accurate are these AbLang (heavy sequences) speed estimates?

They are calculated from specifications rather than measured, and each carries a range — 5,727–15,271 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.

02

What GPU do I need to run AbLang (heavy sequences)?

The smallest card in our catalogue that holds AbLang (heavy sequences) is the Tesla C1080, with 4 GB of memory. It runs the model at Q8_0 using about 1.1 GB, and produces roughly 104 tokens per second. 818 cards in total can run it.

03

How fast is AbLang (heavy sequences) on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 9,544 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 AbLang (heavy sequences) clear that.

04

How much VRAM does AbLang (heavy sequences) 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.

05

Can I run AbLang (heavy sequences) 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 1,778 tokens per second — a comfortable fit.

06

Can I run AbLang (heavy sequences) 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 1,089 tokens per second — a comfortable fit.

07

Can I run AbLang (heavy sequences) 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 1,348 tokens per second — a comfortable fit.

08

Can I run AbLang (heavy sequences) 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 1,599 tokens per second — a comfortable fit.

09

Is AbLang (heavy sequences) open source?

Its weights are published, so AbLang (heavy sequences) 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.

10

How many parameters does AbLang (heavy sequences) have?

AbLang (heavy sequences) has 355M parameters. "The hyperparameters were selected to be similar to those used in the RoBERTa paper (Liu et al., 2019)." Liu et al., 2019 link: https://arxiv.org/pdf/1907.11692.pdf "We begin by training RoBERTa following the BERTLARGE architecture (L = 24, H = 1024, A = 16, 355M 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.

11

Who created AbLang (heavy sequences)?

AbLang (heavy sequences) was published by University of Oxford, based in United Kingdom of Great Britain and Northern Ireland, categorised as academia.

12

When was AbLang (heavy sequences) released?

AbLang (heavy sequences) was published in January 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.

13

What is AbLang (heavy sequences) used for?

AbLang (heavy sequences) works in Biology, and is recorded as handling proteins, Antibody property prediction, Protein or nucleotide language model (pLM/nLM). 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.

14

Where can I download AbLang (heavy sequences)?

The weights for AbLang (heavy sequences) are published, though we do not hold a repository link for it. This site calculates hardware requirements rather than hosting model files.

15

Can I run AbLang (heavy sequences) if it does not fit in my GPU?

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

16

Would two GPUs run AbLang (heavy sequences) faster?

Two cards buy memory rather than speed. That matters for AbLang (heavy sequences) only if one card cannot hold it — 818 can, so a second adds little.

17

Why does the quantisation differ between cards for AbLang (heavy sequences)?

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

Source

Original publication

Record last updated 11 February 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