Big-Little Net TPS calculator

Open weights IBM 77.4M 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 · 477 tok/s

Fastest card

B200

43,798 tok/s · 180 GB

Which GPUs can run Big-Little Net?

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
43,798 tok/s 26,279–70,077 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	0.8 GB	Q8_0	Comfortable
43,798 tok/s 26,279–70,077 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	0.8 GB	Q8_0	Comfortable
34,974 tok/s 20,984–55,958 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	0.8 GB	Q8_0	Comfortable
34,974 tok/s 20,984–55,958 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	0.8 GB	Q8_0	Comfortable
27,971 tok/s 16,782–44,753 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	0.8 GB	Q8_0	Comfortable
26,772 tok/s 16,063–42,835 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	0.8 GB	Q8_0	Comfortable
26,772 tok/s 16,063–42,835 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	0.8 GB	Q8_0	Comfortable
25,622 tok/s 15,373–40,995 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	0.8 GB	Q8_0	Comfortable
22,740 tok/s 13,644–36,383 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	0.8 GB	Q8_0	Comfortable
22,740 tok/s 13,644–36,383 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	0.8 GB	Q8_0	Comfortable
22,740 tok/s 13,644–36,383 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	0.8 GB	Q8_0	Comfortable
21,571 tok/s 12,942–34,513 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	0.8 GB	Q8_0	Comfortable
18,395 tok/s 11,037–29,432 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	0.8 GB	Q8_0	Comfortable
18,395 tok/s 11,037–29,432 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	0.8 GB	Q8_0	Comfortable
18,395 tok/s 11,037–29,432 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	0.8 GB	Q8_0	Comfortable
18,395 tok/s 11,037–29,432 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	0.8 GB	Q8_0	Comfortable
18,395 tok/s 11,037–29,432 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	0.8 GB	Q8_0	Comfortable
14,007 tok/s 8,404–22,411 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	0.8 GB	Q8_0	Comfortable
14,007 tok/s 8,404–22,411 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	0.8 GB	Q8_0	Comfortable
11,672 tok/s 7,003–18,676 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	0.8 GB	Q8_0	Comfortable
11,423 tok/s 6,854–18,277 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	0.8 GB	Q8_0	Comfortable
11,169 tok/s 6,701–17,870 · low confidence	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	0.8 GB	Q8_0	Comfortable
11,169 tok/s 6,701–17,870 · low confidence	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	0.8 GB	Q8_0	Comfortable
11,169 tok/s 6,701–17,870 · low confidence	A800 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Aug 2022	0.8 GB	Q8_0	Comfortable
11,169 tok/s 6,701–17,870 · low confidence	H100 CNX NVIDIA	80 GB	2,040 GB/s	Mar 2023	0.8 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 Chen, Quanfu Fan, Neil Mallinar, Tom Sercu, and Rogerio Feris

What it does

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

Domain: Vision
Task: Image classification, Object recognition

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: 77.4M
Training data: 1,280,000 tokens
Epochs: 110
Batch size: 256

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: 2.5 × 10¹⁷ FLOP
How it was established: 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 Tesla K80

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: Likely
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

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

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

Hardware requirements in practice

Minimum card

Tesla C1080

Memory needed

0.8 GB

Fastest

43,798 tok/s

Big-Little Net is small enough at 77.4M parameters that hardware is rarely the obstacle — 818 of the cards we track can run it, including cards several years old.

The entry point is the Tesla C1080: 4 GB of memory, Q8_0 compression, roughly 477 tokens per second.

The quickest result comes from a B200 at around 43,798 tokens per second — its 8,000 GB/s of bandwidth is what buys that.

What this model is

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

It works in Vision, and is recorded as doing image classification, Object recognition.

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.

What decides the speed

Across every card that can run it, the middle of the range is about 1,229.9 tokens per second, and 818 of them clear the ten tokens per second that roughly matches reading speed.

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.

Training and provenance

Producing it required around 2.5 × 10¹⁷ FLOP of arithmetic, on NVIDIA Tesla K80, which is a statement about the training budget rather than about inference.

It was trained on about 1,280,000 tokens of text.

It is tracked in the underlying dataset for one reason in particular: sOTA improvement.

Step by step

How to choose a GPU for Big-Little Net

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

Every card here has been checked against Big-Little Net — around 0.8 GB at Q8_0. Capacity is the gate — a card either holds it or it does not.
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.
03

Set a quality floor

Compression is what makes Big-Little Net 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 Big-Little Net follows memory bandwidth, not core counts, which is why the B200 tops it at 43,798 tok/s.
05

Read the fit column last

A tight fit runs Big-Little Net 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

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

Answers

Big-Little Net — common questions

When was Big-Little Net released?

Big-Little Net 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 used for?

Big-Little Net works in Vision, and is recorded as handling image classification, Object recognition. Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.

Where can I download Big-Little Net?

The weights for Big-Little Net 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?

Around 2.5 × 10¹⁷ FLOP, on NVIDIA Tesla K80. 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 if it does not fit in my GPU?

Only by offloading, which is usually a false economy: the part in system memory drags the whole thing down. Our figures for Big-Little Net assume it is fully resident.

Would two GPUs run Big-Little Net faster?

Two cards buy memory rather than speed. That matters for Big-Little Net 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?

A larger card holds a more accurate copy. Across the cards that run Big-Little Net, 1 compression levels are used; the floor control above pins it to one.

How accurate are these Big-Little Net 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 26,279–70,077 tok/s on the B200 rather than a single number.

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

The smallest card in our catalogue that holds Big-Little Net is the Tesla C1080, with 4 GB of memory. It runs the model at Q8_0 using about 0.8 GB, and produces roughly 477 tokens per second. 818 cards in total can run it.

How fast is Big-Little Net on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 43,798 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 clear that.

How much VRAM does Big-Little Net need?

About 0.8 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 on a 8 GB GPU?

Yes. A CMP 170HX 8 GB with 8 GB runs it at Q8_0, using about 0.8 GB and generating roughly 8,157 tokens per second — a comfortable fit.

Can I run Big-Little Net on a 12 GB GPU?

Yes. A GeForce RTX 3080 Ti with 12 GB runs it at Q8_0, using about 0.8 GB and generating roughly 4,995 tokens per second — a comfortable fit.

Can I run Big-Little Net on a 16 GB GPU?

Yes. A Tesla V100 SXM2 16 GB with 16 GB runs it at Q8_0, using about 0.8 GB and generating roughly 6,187 tokens per second — a comfortable fit.

Can I run Big-Little Net on a 24 GB GPU?

Yes. A GeForce RTX 5090 D V2 with 24 GB runs it at Q8_0, using about 0.8 GB and generating roughly 7,336 tokens per second — a comfortable fit.

Is Big-Little Net open source?

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

Big-Little Net has 77.4M parameters. Table 2. 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?

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

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