BIDAF TPS calculator

Open weights University of Washington,Allen Institute for AI 2.6M parameters November 2016

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 · 14,178 tok/s

Fastest card

B200

1,303,167 tok/s · 180 GB

Which GPUs can run BIDAF?

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
1,303,167 tok/s 781,900–2,085,068 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	0.7 GB	Q8_0	Comfortable
1,303,167 tok/s 781,900–2,085,068 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	0.7 GB	Q8_0	Comfortable
1,040,612 tok/s 624,367–1,664,979 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	0.7 GB	Q8_0	Comfortable
1,040,612 tok/s 624,367–1,664,979 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	0.7 GB	Q8_0	Comfortable
832,235 tok/s 499,341–1,331,576 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	0.7 GB	Q8_0	Comfortable
796,561 tok/s 477,937–1,274,498 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	0.7 GB	Q8_0	Comfortable
796,561 tok/s 477,937–1,274,498 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	0.7 GB	Q8_0	Comfortable
762,353 tok/s 457,412–1,219,765 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	0.7 GB	Q8_0	Comfortable
676,588 tok/s 405,953–1,082,541 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	0.7 GB	Q8_0	Comfortable
676,588 tok/s 405,953–1,082,541 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	0.7 GB	Q8_0	Comfortable
676,588 tok/s 405,953–1,082,541 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	0.7 GB	Q8_0	Comfortable
641,810 tok/s 385,086–1,026,896 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
547,330 tok/s 328,398–875,729 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
547,330 tok/s 328,398–875,729 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	0.7 GB	Q8_0	Comfortable
547,330 tok/s 328,398–875,729 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
547,330 tok/s 328,398–875,729 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
547,330 tok/s 328,398–875,729 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
416,753 tok/s 250,052–666,805 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	0.7 GB	Q8_0	Comfortable
416,753 tok/s 250,052–666,805 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	0.7 GB	Q8_0	Comfortable
347,294 tok/s 208,376–555,671 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	0.7 GB	Q8_0	Comfortable
339,882 tok/s 203,929–543,812 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	0.7 GB	Q8_0	Comfortable
332,308 tok/s 199,385–531,692 · low confidence	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	0.7 GB	Q8_0	Comfortable
332,308 tok/s 199,385–531,692 · low confidence	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	0.7 GB	Q8_0	Comfortable
332,308 tok/s 199,385–531,692 · low confidence	A800 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Aug 2022	0.7 GB	Q8_0	Comfortable
332,308 tok/s 199,385–531,692 · 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: University of Washington,Allen Institute for AI
Organisation type: Academia,Research collective
Country: United States of America
Published: 5 November 2016
Authors: Minjoon Seo, Aniruddha Kembhavi, Ali Farhadi, Hannaneh Hajishirzi

What it does

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

Domain: Language
Task: Question answering

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: 2.6M
Training data: 879,000 tokens
Epochs: 8

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: 3.5 × 10¹⁸ 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 GeForce GTX TITAN X
Chips used: 8
Chip-hours: 480
Wall-clock time: 60 hours
Power draw: 4.2 kW
Compute cost: $41

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.

Why it is tracked: Highly cited,SOTA improvement
Record confidence: Confident
Citations: 2,246

Sources

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

Reference: Bidirectional Attention Flow for Machine Comprehension
Last updated: 28 November 2025

The extremes

The ten fastest GPUs that run BIDAF

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 BIDAF

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 it takes to run this model

Minimum card

Tesla C1080

Memory needed

0.7 GB

Fastest

1,303,167 tok/s

BIDAF is small enough at 2.6M parameters that hardware is rarely the obstacle — 818 of the cards we track can run it, including cards several years old.

The smallest card that holds it is the Tesla C1080 with 4 GB, running it at Q8_0 and producing around 14,178 tokens per second.

Top of the range is the B200, at roughly 1,303,167 tokens per second thanks to 8,000 GB/s of bandwidth.

Background

BIDAF was published by University of Washington,Allen Institute for AI, in United States of America, in November 2016. academia,Research collective is the category the publisher falls under.

It works in Language, and is recorded as doing question answering.

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.

Reading the throughput figures

The median result is around 36,592.9 tokens per second; 818 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.

Without the attention layout on record, the memory column is an approximation. It is close enough to choose hardware by, and least reliable at long context.

Training and provenance

Producing it required around 3.5 × 10¹⁸ FLOP of arithmetic, on NVIDIA GeForce GTX TITAN X, which is a statement about the training budget rather than about inference.

The training set ran to roughly 879,000 tokens.

Its inclusion criterion is highly cited,SOTA improvement.

Step by step

How to choose a GPU for BIDAF

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

Look at what BIDAF actually needs — around 0.7 GB at Q8_0. No amount of processing power compensates for a card that cannot hold it.
02

Decide how long your conversations run

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

Set a quality floor

Compression is what makes BIDAF 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

Rank by throughput rather than spec sheet

Ranking by tokens per second for BIDAF follows memory bandwidth, not core counts, which is why the B200 tops it at 1,303,167 tok/s.
05

Look at the headroom, not just the fit

A tight fit runs BIDAF but leaves nothing spare for a longer conversation; comfortable has headroom. If you expect to grow the context, buy for comfortable.
06

Check the card from the other side

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

Answers

BIDAF — common questions

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

Would two GPUs run BIDAF faster?

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

Why does the quantisation differ between cards for BIDAF?

Because capacity varies, so does how hard BIDAF has to be squeezed — 1 distinct levels appear in the table above. Set a minimum quality to compare at one.

How accurate are these BIDAF speed estimates?

These are estimates with real error bars. The fastest result here, 781,900–2,085,068 tok/s on the B200, could reasonably land anywhere in its published range depending on which runtime you use.

What GPU do I need to run BIDAF?

The smallest card in our catalogue that holds BIDAF is the Tesla C1080, with 4 GB of memory. It runs the model at Q8_0 using about 0.7 GB, and produces roughly 14,178 tokens per second. 818 cards in total can run it.

How fast is BIDAF on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 1,303,167 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 BIDAF clear that.

How much VRAM does BIDAF 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 BIDAF 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 242,715 tokens per second — a comfortable fit.

Can I run BIDAF 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 148,626 tokens per second — a comfortable fit.

Can I run BIDAF 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 184,072 tokens per second — a comfortable fit.

Can I run BIDAF 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 218,281 tokens per second — a comfortable fit.

Is BIDAF open source?

Its weights are published, so BIDAF 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 BIDAF have?

BIDAF has 2.6M parameters. There are two similar models described in sections "Models details" citation from the paper about model for SQuAD "The model has about 2.6 million parameters" citation about model for cloze test "The model architecture used for this task is very similar to that for SQuAD (Section 4) with only a few small changes to adapt it to the cloze test. ". 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 BIDAF?

BIDAF was published by University of Washington,Allen Institute for AI, based in United States of America, categorised as academia,Research collective.

When was BIDAF released?

BIDAF was published in November 2016. 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 BIDAF used for?

BIDAF works in Language, and is recorded as handling question answering. Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.

Where can I download BIDAF?

The weights for BIDAF 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 BIDAF?

Around 3.5 × 10¹⁸ FLOP, on NVIDIA GeForce GTX TITAN X. 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 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