ProtBFN

Closed weights InstaDeep 650M parameters September 2024

No estimate

No hardware requirements for this model

The weights for this model have not been published, so it cannot be downloaded or run on your own hardware at any size. It is reachable only through its provider, and no graphics card changes that.

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
InstaDeep
Organisation type
Industry
Country
United Kingdom of Great Britain and Northern Ireland
Published
24 September 2024
Authors
Timothy Atkinson, Thomas D. Barrett, Scott Cameron, Bora Guloglu, Matthew Greenig, Louis Robinson, Alex Graves, Liviu Copoiu, Alexandre Laterre

What it does

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

Domain
Biology
Task
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
650M
Training data
tokens

Total tokens = 71,000,000 sequences × 300 residues = 21,300,000,000 tokens (2.13×10¹⁰) Alternative calculation with 256 residues: 71,000,000 sequences × 256 residues = 18,176,000,000 tokens (1.82×10¹⁰) Final estimate: 2.1×10¹⁰ 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
3.9 × 10²² FLOP

1. Hardware setup: - ProtBFN: 256 TPU v4 chips (2.75e14 FLOP/s per chip) - AbBFN: 128 TPU v4 chips (2.75e14 FLOP/s per chip) 2. Training duration: - ProtBFN: 14 days = 1,209,600 seconds - AbBFN: 4 days = 345,600 seconds 3. Utilization rate: 40% 4. Final calculation: - ProtBFN: 2.75e14 × 256 × 1,209,600 × 0.4 = 3.406e22 FLOPs - AbBFN: 2.75e14 × 128 × 345,600 × 0.4 = 4.866e21 FLOPs - Total: 3.406e22 + 4.866e21 = 3.9e22 FLOPs

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
Google TPU v4
Chips used
256
Wall-clock time
432 hours (18 days)
Power draw
171.4 kW

How it is classified

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

Record confidence
Confident
Citations
16

Sources

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

Reference
Protein Sequence Modelling with Bayesian Flow Networks
Last updated
1 January 2026

What the numbers mean

Background

ProtBFN was published by InstaDeep, in United Kingdom of Great Britain and Northern Ireland, in September 2024. industry is the category the publisher falls under.

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

Because the weights are not available, none of the hardware figures elsewhere on this site apply to it.

What went into building it

Producing it required around 3.9 × 10²² FLOP of arithmetic, on Google TPU v4, which is a statement about the training budget rather than about inference.

Answers

ProtBFN — common questions

01

How many parameters does ProtBFN have?

ProtBFN has 650M 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.

02

Who created ProtBFN?

ProtBFN was published by InstaDeep, based in United Kingdom of Great Britain and Northern Ireland, categorised as industry.

03

When was ProtBFN released?

ProtBFN was published in September 2024. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.

04

What is ProtBFN used for?

ProtBFN works in Biology, and is recorded as handling 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.

05

How much compute was used to train ProtBFN?

Around 3.9 × 10²² FLOP, on Google TPU v4. 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.

06

What GPU do I need to run ProtBFN?

None. ProtBFN is a closed model — its weights were never published, so it cannot be downloaded or run on your own hardware at any price. It is reachable only through its provider.

07

Is ProtBFN open source?

The licensing for ProtBFN was never recorded in our source data. We treat unstated licensing as closed, because an unrecorded licence is not one to rely on.

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.