Big-Little Net (speech) TPS calculator

Open weights IBM 3.3M parameters July 2018

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 · 11,104 tok/s

Fastest card

B200

1,020,553 tok/s · 180 GB

Which GPUs can run Big-Little Net (speech)?

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
1,020,553 tok/s 612,332–1,632,884 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	0.7 GB	Q8_0	Comfortable
1,020,553 tok/s 612,332–1,632,884 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	0.7 GB	Q8_0	Comfortable
814,937 tok/s 488,962–1,303,899 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	0.7 GB	Q8_0	Comfortable
814,937 tok/s 488,962–1,303,899 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	0.7 GB	Q8_0	Comfortable
651,751 tok/s 391,050–1,042,801 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	0.7 GB	Q8_0	Comfortable
623,813 tok/s 374,288–998,101 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	0.7 GB	Q8_0	Comfortable
623,813 tok/s 374,288–998,101 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	0.7 GB	Q8_0	Comfortable
597,023 tok/s 358,214–955,237 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	0.7 GB	Q8_0	Comfortable
529,858 tok/s 317,915–847,773 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	0.7 GB	Q8_0	Comfortable
529,858 tok/s 317,915–847,773 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	0.7 GB	Q8_0	Comfortable
529,858 tok/s 317,915–847,773 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	0.7 GB	Q8_0	Comfortable
502,622 tok/s 301,573–804,196 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
428,632 tok/s 257,179–685,811 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
428,632 tok/s 257,179–685,811 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	0.7 GB	Q8_0	Comfortable
428,632 tok/s 257,179–685,811 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
428,632 tok/s 257,179–685,811 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
428,632 tok/s 257,179–685,811 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
326,373 tok/s 195,824–522,196 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	0.7 GB	Q8_0	Comfortable
326,373 tok/s 195,824–522,196 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	0.7 GB	Q8_0	Comfortable
271,977 tok/s 163,186–435,164 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	0.7 GB	Q8_0	Comfortable
266,173 tok/s 159,704–425,877 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	0.7 GB	Q8_0	Comfortable
260,241 tok/s 156,145–416,386 · low confidence	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	0.7 GB	Q8_0	Comfortable
260,241 tok/s 156,145–416,386 · low confidence	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	0.7 GB	Q8_0	Comfortable
260,241 tok/s 156,145–416,386 · low confidence	A800 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Aug 2022	0.7 GB	Q8_0	Comfortable
260,241 tok/s 156,145–416,386 · 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: IBM
Organisation type: Industry
Country: United States of America
Published: 10 July 2018
Authors: Chun-Fu (Richard) Chen, Quanfu Fan, Neil Mallinar, Tom Sercu, Rogerio Feris

What it does

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

Domain: Speech
Task: Speech recognition (ASR)

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: 3.3M
Training data: 720,000,000 tokens
Epochs: 16

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: 4.3 × 10¹⁷ FLOP
How it was established: Operation counting

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: SOTA improvement
Record confidence: Speculative
Citations: 101

Sources

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

Reference: Big-Little Net: An Efficient Multi-Scale Feature Representation for Visual and Speech Recognition
Last updated: 25 May 2026

The extremes

The ten fastest GPUs that run Big-Little Net (speech)

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 Big-Little Net (speech)

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

1,020,553 tok/s

Big-Little Net (speech) is small enough at 3.3M 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 11,104 tokens per second.

A B200 is the fastest we calculate for it: about 1,020,553 tokens per second, from 8,000 GB/s of memory bandwidth.

Where it came from

Big-Little Net (speech) was published by IBM, in United States of America, in July 2018. The organisation is categorised as industry.

It works in Speech, and is recorded as doing speech recognition (ASR).

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 28,657.1 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.

What went into building it

Producing it required around 4.3 × 10¹⁷ FLOP of arithmetic, which is a statement about the training budget rather than about inference.

It was trained on about 720,000,000 tokens of text.

The reason it appears in this catalogue at all is sOTA improvement.

Step by step

How to choose a GPU for Big-Little Net (speech)

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

01

Check what it needs before anything else

The table lists every card that can hold Big-Little Net (speech) — 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

Longer conversations cost memory on top of what the weights need. Move the slider to your real working length before trusting any row for Big-Little Net (speech).
03

Decide how much compression you will accept

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

Sort by speed

The speed ordering for Big-Little Net (speech) is effectively an ordering by memory bandwidth, which is why the B200 tops it at 1,020,553 tok/s.
05

Read the fit column last

The fit column separates cards that just manage Big-Little Net (speech) from those with room to spare. Buy for the second if the context might grow.
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 Big-Little Net (speech).

Answers

Big-Little Net (speech) — common questions

What GPU do I need to run Big-Little Net (speech)?

The smallest card in our catalogue that holds Big-Little Net (speech) is the Tesla C1080, with 4 GB of memory. It runs the model at Q8_0 using about 0.7 GB, and produces roughly 11,104 tokens per second. 818 cards in total can run it.

How fast is Big-Little Net (speech) on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 1,020,553 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 Big-Little Net (speech) clear that.

How much VRAM does Big-Little Net (speech) 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 Big-Little Net (speech) 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 190,078 tokens per second — a comfortable fit.

Can I run Big-Little Net (speech) 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 116,394 tokens per second — a comfortable fit.

Can I run Big-Little Net (speech) 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 144,153 tokens per second — a comfortable fit.

Can I run Big-Little Net (speech) 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 170,943 tokens per second — a comfortable fit.

Is Big-Little Net (speech) open source?

Its weights are published, so Big-Little Net (speech) 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 Big-Little Net (speech) have?

Big-Little Net (speech) has 3.3M parameters. table 3. 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 Big-Little Net (speech)?

Big-Little Net (speech) was published by IBM, based in United States of America, categorised as industry.

When was Big-Little Net (speech) released?

Big-Little Net (speech) was published in July 2018. 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 Big-Little Net (speech) used for?

Big-Little Net (speech) works in Speech, and is recorded as handling speech recognition (ASR). These are the areas it was designed around; they describe intent rather than a hard boundary.

Where can I download Big-Little Net (speech)?

The weights for Big-Little Net (speech) 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 Big-Little Net (speech)?

Around 4.3 × 10¹⁷ FLOP. 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 Big-Little Net (speech) if it does not fit in my GPU?

It can be split between the card and system memory, but Big-Little Net (speech) generates painfully slowly that way. Nothing on this page assumes offloading.

Would two GPUs run Big-Little Net (speech) faster?

Two cards buy memory rather than speed. That matters for Big-Little Net (speech) only if one card cannot hold it — 818 can, so a second adds little.

Why does the quantisation differ between cards for Big-Little Net (speech)?

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

How accurate are these Big-Little Net (speech) speed estimates?

They are calculated from specifications rather than measured, and each carries a range — 612,332–1,632,884 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.

Source

Original publication

Record last updated 25 May 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