ESM1b TPS calculator

Open weights Facebook AI Research,New York University (NYU) 652.4M parameters December 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 · 56.5 tok/s

Fastest card

B200

5,193 tok/s · 180 GB

Which GPUs can run ESM1b?

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
5,193 tok/s 3,116–8,310 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	1.4 GB	Q8_0	Comfortable
5,193 tok/s 3,116–8,310 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	1.4 GB	Q8_0	Comfortable
4,147 tok/s 2,488–6,635 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	1.4 GB	Q8_0	Comfortable
4,147 tok/s 2,488–6,635 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	1.4 GB	Q8_0	Comfortable
3,317 tok/s 1,990–5,307 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	1.4 GB	Q8_0	Comfortable
3,175 tok/s 1,905–5,079 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	1.4 GB	Q8_0	Comfortable
3,175 tok/s 1,905–5,079 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	1.4 GB	Q8_0	Comfortable
3,038 tok/s 1,823–4,861 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	1.4 GB	Q8_0	Comfortable
2,696 tok/s 1,618–4,314 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	1.4 GB	Q8_0	Comfortable
2,696 tok/s 1,618–4,314 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	1.4 GB	Q8_0	Comfortable
2,696 tok/s 1,618–4,314 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	1.4 GB	Q8_0	Comfortable
2,558 tok/s 1,535–4,092 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	1.4 GB	Q8_0	Comfortable
2,181 tok/s 1,309–3,490 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	1.4 GB	Q8_0	Comfortable
2,181 tok/s 1,309–3,490 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	1.4 GB	Q8_0	Comfortable
2,181 tok/s 1,309–3,490 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	1.4 GB	Q8_0	Comfortable
2,181 tok/s 1,309–3,490 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	1.4 GB	Q8_0	Comfortable
2,181 tok/s 1,309–3,490 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	1.4 GB	Q8_0	Comfortable
1,661 tok/s 997–2,657 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	1.4 GB	Q8_0	Comfortable
1,661 tok/s 997–2,657 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	1.4 GB	Q8_0	Comfortable
1,384 tok/s 830–2,215 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	1.4 GB	Q8_0	Comfortable
1,355 tok/s 813–2,167 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	1.4 GB	Q8_0	Comfortable
1,324 tok/s 795–2,119 · low confidence	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	1.4 GB	Q8_0	Comfortable
1,324 tok/s 795–2,119 · low confidence	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	1.4 GB	Q8_0	Comfortable
1,324 tok/s 795–2,119 · low confidence	A800 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Aug 2022	1.4 GB	Q8_0	Comfortable
1,324 tok/s 795–2,119 · 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: Facebook AI Research,New York University (NYU)
Organisation type: Industry,Academia
Country: United States of America, France
Published: 15 December 2020
Authors: Alexander Rives, Joshua Meier, Tom Sercu, Siddharth Goyal, Zeming Lin, Jason Liu, Demi Guo, Myle Ott, C. Lawrence Zitnick, Jerry Ma, and Rob Fergus

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)

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: 652.4M
Training data: 27,750,400,000 tokens
Epochs: 56

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: 5.1 × 10²¹ FLOP
How it was established: Hardware,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: NVIDIA V100
Chips used: 128
Power draw: 78.0 kW
Compute cost: $924

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: Highly cited,SOTA improvement
Record confidence: Confident
Citations: 2,637

Sources

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

Reference: Biological structure and function emerge from scaling unsupervised learning to 250 million protein sequences
Last updated: 1 January 2026

The extremes

The ten fastest GPUs that run ESM1b

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 ESM1b

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

1.4 GB

Fastest

5,193 tok/s

ESM1b is small enough at 652.4M 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 56.5 tokens per second.

The quickest result comes from a B200 at around 5,193 tokens per second — its 8,000 GB/s of bandwidth is what buys that.

What this model is

ESM1b was published by Facebook AI Research,New York University (NYU), in United States of America, in December 2020. The organisation is categorised as industry,Academia.

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

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

What decides the speed

The median result is around 145.8 tokens per second; 809 cards produce text faster than most people read it.

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.

Memory here is estimated from size rather than computed from the architecture, which is not recorded for this model — the numbers are indicative rather than exact.

What went into building it

The training run consumed about 5.1 × 10²¹ FLOP, on NVIDIA V100. That figure describes the cost of creating it and has no bearing on how quickly it generates text.

It was trained on about 27,750,400,000 tokens of text.

It is tracked in the underlying dataset for one reason in particular: highly cited,SOTA improvement.

Step by step

How to choose a GPU for ESM1b

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

The table lists every card that can hold ESM1b — around 1.4 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

Set the context to what you will actually use. The cache grows with the conversation, and it is the usual reason ESM1b stops fitting a card that seemed fine.
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 ESM1b by squeezing it further than you would want.
04

Rank by throughput rather than spec sheet

The speed ordering for ESM1b is effectively an ordering by memory bandwidth, which is why the B200 tops it at 5,193 tok/s.
05

Look at the headroom, not just the fit

The fit column separates cards that just manage ESM1b from those with room to spare. Buy for the second if the context might grow.
06

Open the card you have settled on

Each card page repeats this sweep for every model we hold. It answers what else the hardware is good for, beyond ESM1b.

Answers

ESM1b — common questions

What GPU do I need to run ESM1b?

The smallest card in our catalogue that holds ESM1b is the Tesla C1080, with 4 GB of memory. It runs the model at Q8_0 using about 1.4 GB, and produces roughly 56.5 tokens per second. 818 cards in total can run it.

How fast is ESM1b on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 5,193 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 ESM1b clear that.

How much VRAM does ESM1b 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 ESM1b 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 967 tokens per second — a comfortable fit.

Can I run ESM1b 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 592 tokens per second — a comfortable fit.

Can I run ESM1b 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 734 tokens per second — a comfortable fit.

Can I run ESM1b 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 870 tokens per second — a comfortable fit.

Is ESM1b open source?

Its weights are published, so ESM1b 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 ESM1b have?

ESM1b has 652.4M parameters. See Table 9. 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 ESM1b?

ESM1b was published by Facebook AI Research,New York University (NYU), based in United States of America, categorised as industry,Academia.

When was ESM1b released?

ESM1b was published in December 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 ESM1b used for?

ESM1b 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 ESM1b?

The weights for ESM1b 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 ESM1b?

Around 5.1 × 10²¹ 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 ESM1b if it does not fit in my GPU?

It can be split between the card and system memory, but ESM1b generates painfully slowly that way. Nothing on this page assumes offloading.

Would two GPUs run ESM1b faster?

A second card roughly doubles the memory available but not the generation rate. With 818 cards already able to run ESM1b alone, the case for pairing is weak.

Why does the quantisation differ between cards for ESM1b?

Each card is shown running the least-compressed copy it can hold, and ESM1b appears at 1 different compression levels across the cards that fit it. Bigger cards get the more accurate version.

How accurate are these ESM1b speed estimates?

These are estimates with real error bars. The fastest result here, 3,116–8,310 tok/s on the B200, could reasonably land anywhere in its published range depending on which runtime you use.

Source

Original publication

Record last updated 1 January 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