HiFi - NN TPS calculator

Open weights Basecamp Research,Technical University of Munich,Molecular Institute of Biology,Microsoft Research 3M parameters December 2023

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 · 12,288 tok/s

Fastest card

B200

1,129,412 tok/s · 180 GB

Which GPUs can run HiFi - NN?

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
1,129,412 tok/s 677,647–1,807,059 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	0.7 GB	Q8_0	Comfortable
1,129,412 tok/s 677,647–1,807,059 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	0.7 GB	Q8_0	Comfortable
901,864 tok/s 541,118–1,442,982 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	0.7 GB	Q8_0	Comfortable
901,864 tok/s 541,118–1,442,982 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	0.7 GB	Q8_0	Comfortable
721,271 tok/s 432,762–1,154,033 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	0.7 GB	Q8_0	Comfortable
690,353 tok/s 414,212–1,104,565 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	0.7 GB	Q8_0	Comfortable
690,353 tok/s 414,212–1,104,565 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	0.7 GB	Q8_0	Comfortable
660,706 tok/s 396,424–1,057,129 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	0.7 GB	Q8_0	Comfortable
586,376 tok/s 351,826–938,202 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	0.7 GB	Q8_0	Comfortable
586,376 tok/s 351,826–938,202 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	0.7 GB	Q8_0	Comfortable
586,376 tok/s 351,826–938,202 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	0.7 GB	Q8_0	Comfortable
556,235 tok/s 333,741–889,976 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
474,353 tok/s 284,612–758,965 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
474,353 tok/s 284,612–758,965 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	0.7 GB	Q8_0	Comfortable
474,353 tok/s 284,612–758,965 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
474,353 tok/s 284,612–758,965 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
474,353 tok/s 284,612–758,965 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
361,186 tok/s 216,712–577,897 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	0.7 GB	Q8_0	Comfortable
361,186 tok/s 216,712–577,897 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	0.7 GB	Q8_0	Comfortable
300,988 tok/s 180,593–481,581 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	0.7 GB	Q8_0	Comfortable
294,565 tok/s 176,739–471,304 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	0.7 GB	Q8_0	Comfortable
288,000 tok/s 172,800–460,800 · low confidence	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	0.7 GB	Q8_0	Comfortable
288,000 tok/s 172,800–460,800 · low confidence	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	0.7 GB	Q8_0	Comfortable
288,000 tok/s 172,800–460,800 · low confidence	A800 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Aug 2022	0.7 GB	Q8_0	Comfortable
288,000 tok/s 172,800–460,800 · 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: Basecamp Research,Technical University of Munich,Molecular Institute of Biology,Microsoft Research
Organisation type: Industry,Academia,Academia,Industry
Country: United Kingdom of Great Britain and Northern Ireland, Germany, Spain, United States of America
Published: 19 December 2023
Authors: Gavin Ayres

What it does

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

Domain: Biology
Task: Enzyme function prediction

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: 3M
Training data: tokens

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: 8
Power draw: 6.3 kW

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

How it is classified

Labels the source dataset applies when tracking notable models, and how confident it is in the entry.

Record confidence: Confident

Sources

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

Reference: Breakthrough in Functional Annotation with HiFi-NN
Last updated: 28 November 2025

The extremes

The ten fastest GPUs that run HiFi - NN

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 HiFi - NN

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

1,129,412 tok/s

HiFi - NN is small enough at 3M 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 12,288 tokens per second.

Top of the range is the B200, at roughly 1,129,412 tokens per second thanks to 8,000 GB/s of bandwidth.

Where it came from

HiFi - NN was published by Basecamp Research,Technical University of Munich,Molecular Institute of Biology,Microsoft Research, in United Kingdom of Great Britain and Northern Ireland, in December 2023. It comes out of industry,Academia,Academia,Industry.

It works in Biology, and is recorded as doing enzyme function prediction.

The weights are published, so it can be downloaded and run on your own hardware indefinitely, offline, with no account attached.

Understanding the speeds

Across every card that can run it, the middle of the range is about 31,713.9 tokens per second, and 818 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.

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.

Step by step

How to choose a GPU for HiFi - NN

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 HiFi - NN — around 0.7 GB at Q8_0. That figure, not the card's headline performance, is what decides whether it runs.
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 HiFi - NN can slip off a card that handles short questions easily.
03

Choose how far you will compress it

The quantisation column varies by card, because a bigger card holds a more accurate copy of HiFi - NN — Q8_0 on the smallest card that fits. Set a floor to hold the comparison at one level.
04

Rank by throughput rather than spec sheet

Ranking by tokens per second for HiFi - NN follows memory bandwidth, not core counts, which is why the B200 tops it at 1,129,412 tok/s.
05

Check the fit verdict before buying

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

See what else that card runs

Following a card through to its own page shows every other model it can hold, which is the question that follows once HiFi - NN is settled.

Answers

HiFi - NN — common questions

Would two GPUs run HiFi - NN faster?

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

Why does the quantisation differ between cards for HiFi - NN?

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

How accurate are these HiFi - NN speed estimates?

Every figure is derived from memory bandwidth and model size, not benchmarked. That is why each is published as a range such as 677,647–1,807,059 tok/s on the B200 rather than a single number.

What GPU do I need to run HiFi - NN?

The smallest card in our catalogue that holds HiFi - NN is the Tesla C1080, with 4 GB of memory. It runs the model at Q8_0 using about 0.7 GB, and produces roughly 12,288 tokens per second. 818 cards in total can run it.

How fast is HiFi - NN on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 1,129,412 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 HiFi - NN clear that.

How much VRAM does HiFi - NN 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 HiFi - NN 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 210,353 tokens per second — a comfortable fit.

Can I run HiFi - NN 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 128,809 tokens per second — a comfortable fit.

Can I run HiFi - NN 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 159,529 tokens per second — a comfortable fit.

Can I run HiFi - NN 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 189,176 tokens per second — a comfortable fit.

Is HiFi - NN open source?

Its weights are published, so HiFi - NN 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 HiFi - NN have?

HiFi - NN has 3M parameters. "The model boasts over 3M 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 HiFi - NN?

HiFi - NN was published by Basecamp Research,Technical University of Munich,Molecular Institute of Biology,Microsoft Research, based in United Kingdom of Great Britain and Northern Ireland, categorised as industry,Academia,Academia,Industry.

When was HiFi - NN released?

HiFi - NN was published in December 2023. 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 HiFi - NN used for?

HiFi - NN works in Biology, and is recorded as handling enzyme function prediction. These are the areas it was designed around; they describe intent rather than a hard boundary.

Where can I download HiFi - NN?

The weights for HiFi - NN are published, though we do not hold a repository link for it. This site calculates hardware requirements rather than hosting model files.

Can I run HiFi - NN if it does not fit in my GPU?

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

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