MobileBERT TPS calculator

Open weights Carnegie Mellon University (CMU),Google Brain 25.3M parameters April 2020

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 · 1,457 tok/s

Fastest card

B200

133,922 tok/s · 180 GB

Which GPUs can run MobileBERT?

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
133,922 tok/s 80,353–214,276 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	0.7 GB	Q8_0	Comfortable
133,922 tok/s 80,353–214,276 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	0.7 GB	Q8_0	Comfortable
106,940 tok/s 64,164–171,105 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	0.7 GB	Q8_0	Comfortable
106,940 tok/s 64,164–171,105 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	0.7 GB	Q8_0	Comfortable
85,526 tok/s 51,316–136,842 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	0.7 GB	Q8_0	Comfortable
81,860 tok/s 49,116–130,976 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	0.7 GB	Q8_0	Comfortable
81,860 tok/s 49,116–130,976 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	0.7 GB	Q8_0	Comfortable
78,345 tok/s 47,007–125,351 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	0.7 GB	Q8_0	Comfortable
69,531 tok/s 41,718–111,249 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	0.7 GB	Q8_0	Comfortable
69,531 tok/s 41,718–111,249 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	0.7 GB	Q8_0	Comfortable
69,531 tok/s 41,718–111,249 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	0.7 GB	Q8_0	Comfortable
65,957 tok/s 39,574–105,531 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
56,247 tok/s 33,748–89,996 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
56,247 tok/s 33,748–89,996 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	0.7 GB	Q8_0	Comfortable
56,247 tok/s 33,748–89,996 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
56,247 tok/s 33,748–89,996 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
56,247 tok/s 33,748–89,996 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
42,828 tok/s 25,697–68,525 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	0.7 GB	Q8_0	Comfortable
42,828 tok/s 25,697–68,525 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	0.7 GB	Q8_0	Comfortable
35,690 tok/s 21,414–57,104 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	0.7 GB	Q8_0	Comfortable
34,929 tok/s 20,957–55,886 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	0.7 GB	Q8_0	Comfortable
34,150 tok/s 20,490–54,640 · low confidence	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	0.7 GB	Q8_0	Comfortable
34,150 tok/s 20,490–54,640 · low confidence	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	0.7 GB	Q8_0	Comfortable
34,150 tok/s 20,490–54,640 · low confidence	A800 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Aug 2022	0.7 GB	Q8_0	Comfortable
34,150 tok/s 20,490–54,640 · 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: Carnegie Mellon University (CMU),Google Brain
Organisation type: Academia,Industry
Country: United States of America
Published: 6 April 2020
Authors: Zhiqing Sun, Hongkun Yu, Xiaodan Song, Renjie Liu, Yiming Yang, Denny Zhou

What it does

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

Domain: Language
Task: Text autocompletion, Language modeling/generation

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: 25.3M
Training data: tokens

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: Unreleased
Hugging Face: google

How it is classified

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

Citations: 1,011

Sources

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

Reference: MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices
Last updated: 25 May 2026

The extremes

The ten fastest GPUs that run MobileBERT

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 MobileBERT

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

133,922 tok/s

MobileBERT is small enough at 25.3M 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 1,457 tokens per second.

Top of the range is the B200, at roughly 133,922 tokens per second thanks to 8,000 GB/s of bandwidth.

About this model

MobileBERT was published by Carnegie Mellon University (CMU),Google Brain, in United States of America, in April 2020. It comes out of academia,Industry.

It works in Language, and is recorded as doing text autocompletion, Language modeling/generation.

Because its weights were released, nothing about running it depends on a provider staying available — it is yours once downloaded. It is published under the google organisation on Hugging Face.

How fast it runs, and why

Across every card that can run it, the middle of the range is about 3,760.5 tokens per second, and 818 of them clear the ten tokens per second that roughly matches reading speed.

Every weight participates in every token here, so bandwidth is the whole story: the ranking below is effectively a ranking of memory throughput.

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.

Step by step

How to choose a GPU for MobileBERT

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

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

Match the context to your actual use

Set the context to what you will actually use. The cache grows with the conversation, and it is the usual reason MobileBERT stops fitting a card that seemed fine.
03

Set a quality floor

The quantisation column varies by card, because a bigger card holds a more accurate copy of MobileBERT — 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 MobileBERT follows memory bandwidth, not core counts, which is why the B200 tops it at 133,922 tok/s.
05

Check the fit verdict before buying

The fit column separates cards that just manage MobileBERT 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 MobileBERT.

Answers

MobileBERT — common questions

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

Would two GPUs run MobileBERT faster?

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

Why does the quantisation differ between cards for MobileBERT?

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

How accurate are these MobileBERT speed estimates?

They are calculated from specifications rather than measured, and each carries a range — 80,353–214,276 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.

What GPU do I need to run MobileBERT?

The smallest card in our catalogue that holds MobileBERT is the Tesla C1080, with 4 GB of memory. It runs the model at Q8_0 using about 0.7 GB, and produces roughly 1,457 tokens per second. 818 cards in total can run it.

How fast is MobileBERT on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 133,922 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 MobileBERT clear that.

How much VRAM does MobileBERT 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 MobileBERT 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 24,943 tokens per second — a comfortable fit.

Can I run MobileBERT 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 15,274 tokens per second — a comfortable fit.

Can I run MobileBERT 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 18,917 tokens per second — a comfortable fit.

Can I run MobileBERT 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 22,432 tokens per second — a comfortable fit.

Is MobileBERT open source?

Its weights are published, so MobileBERT 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 MobileBERT have?

MobileBERT has 25.3M parameters. Rados. 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 MobileBERT?

MobileBERT was published by Carnegie Mellon University (CMU),Google Brain, based in United States of America, categorised as academia,Industry.

When was MobileBERT released?

MobileBERT was published in April 2020. 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 MobileBERT used for?

MobileBERT works in Language, and is recorded as handling text autocompletion, Language modeling/generation. Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.

Where can I download MobileBERT?

Its weights are published under the google organisation on Hugging Face. We do not host model files — this site calculates what hardware is needed to run them.

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