mBART-50 TPS calculator

Open weights Facebook AI 610M parameters August 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 · 60.4 tok/s

Fastest card

B200

5,554 tok/s · 180 GB

Which GPUs can run mBART-50?

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
5,554 tok/s 3,333–8,887 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	1.4 GB	Q8_0	Comfortable
5,554 tok/s 3,333–8,887 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	1.4 GB	Q8_0	Comfortable
4,435 tok/s 2,661–7,097 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	1.4 GB	Q8_0	Comfortable
4,435 tok/s 2,661–7,097 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	1.4 GB	Q8_0	Comfortable
3,547 tok/s 2,128–5,676 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	1.4 GB	Q8_0	Comfortable
3,395 tok/s 2,037–5,432 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	1.4 GB	Q8_0	Comfortable
3,395 tok/s 2,037–5,432 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	1.4 GB	Q8_0	Comfortable
3,249 tok/s 1,950–5,199 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	1.4 GB	Q8_0	Comfortable
2,884 tok/s 1,730–4,614 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	1.4 GB	Q8_0	Comfortable
2,884 tok/s 1,730–4,614 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	1.4 GB	Q8_0	Comfortable
2,884 tok/s 1,730–4,614 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	1.4 GB	Q8_0	Comfortable
2,736 tok/s 1,641–4,377 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	1.4 GB	Q8_0	Comfortable
2,333 tok/s 1,400–3,733 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	1.4 GB	Q8_0	Comfortable
2,333 tok/s 1,400–3,733 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	1.4 GB	Q8_0	Comfortable
2,333 tok/s 1,400–3,733 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	1.4 GB	Q8_0	Comfortable
2,333 tok/s 1,400–3,733 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	1.4 GB	Q8_0	Comfortable
2,333 tok/s 1,400–3,733 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	1.4 GB	Q8_0	Comfortable
1,776 tok/s 1,066–2,842 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	1.4 GB	Q8_0	Comfortable
1,776 tok/s 1,066–2,842 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	1.4 GB	Q8_0	Comfortable
1,480 tok/s 888–2,368 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	1.4 GB	Q8_0	Comfortable
1,449 tok/s 869–2,318 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	1.4 GB	Q8_0	Comfortable
1,416 tok/s 850–2,266 · low confidence	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	1.4 GB	Q8_0	Comfortable
1,416 tok/s 850–2,266 · low confidence	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	1.4 GB	Q8_0	Comfortable
1,416 tok/s 850–2,266 · low confidence	A800 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Aug 2022	1.4 GB	Q8_0	Comfortable
1,416 tok/s 850–2,266 · low confidence	H100 CNX NVIDIA	80 GB	2,040 GB/s	Mar 2023	1.4 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: Facebook AI
Organisation type: Industry
Country: United States of America
Published: 2 August 2020
Authors: Yuqing Tang, Chau Tran, Xian Li, Peng-Jen Chen, Naman Goyal, Vishrav Chaudhary, Jiatao Gu, Angela Fan

What it does

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

Domain: Language
Task: Translation

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: 610M
Training data: 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: 1.5 × 10²² 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 V100
Chips used: 256
Chip-hours: 107,520
Wall-clock time: 420 hours (17.5 days)
Power draw: 156.5 kW

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)

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: 552

Sources

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

Reference: Multilingual Translation with Extensible Multilingual Pretraining and Finetuning
Last updated: 25 May 2026

The extremes

The ten fastest GPUs that run mBART-50

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 mBART-50

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

1.4 GB

Fastest

5,554 tok/s

mBART-50 is small enough at 610M 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 60.4 tokens per second.

A B200 is the fastest we calculate for it: about 5,554 tokens per second, from 8,000 GB/s of memory bandwidth.

Where it came from

mBART-50 was published by Facebook AI, in United States of America, in August 2020. industry is the category the publisher falls under.

It works in Language, and is recorded as doing translation.

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.

Understanding the speeds

Across every card that can run it, the middle of the range is about 156.0 tokens per second, and 809 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.

Its internal architecture is not on file, so memory is approximated from the parameter count and marked accordingly. Expect the real figure to differ, more so at long context.

Training and provenance

The training run consumed about 1.5 × 10²² FLOP, on NVIDIA V100. That figure describes the cost of creating it and has no bearing on how quickly it generates text.

Step by step

How to choose a GPU for mBART-50

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

Every card here has been checked against mBART-50 — around 1.4 GB at Q8_0. Capacity is the gate — a card either holds it or it does not.
02

Decide how long your conversations run

Longer conversations cost memory on top of what the weights need. Move the slider to your real working length before trusting any row for mBART-50.
03

Set a quality floor

Compression is what makes mBART-50 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 mBART-50 follows memory bandwidth, not core counts, which is why the B200 tops it at 5,554 tok/s.
05

Read the fit column last

Tight means mBART-50 loads and works, with no room to raise the context later. Comfortable means you can. The difference matters more than a few tokens per second.
06

Open the card you have settled on

Following a card through to its own page shows every other model it can hold, which is the question that follows once mBART-50 is settled.

Answers

mBART-50 — common questions

Can I run mBART-50 on a 24 GB GPU?

Yes. A GeForce RTX 5090 D V2 with 24 GB runs it at Q8_0, using about 1.4 GB and generating roughly 930 tokens per second — a comfortable fit.

Is mBART-50 open source?

Its weights are published, so mBART-50 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 mBART-50 have?

mBART-50 has 610M parameters. 610M from https://github.com/facebookresearch/fairseq/tree/main/examples/mbart. 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 mBART-50?

mBART-50 was published by Facebook AI, based in United States of America, categorised as industry.

When was mBART-50 released?

mBART-50 was published in August 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 mBART-50 used for?

mBART-50 works in Language, and is recorded as handling translation. Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.

Where can I download mBART-50?

The weights for mBART-50 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 mBART-50?

Around 1.5 × 10²² FLOP, on NVIDIA V100. 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 mBART-50 if it does not fit in my GPU?

It can be split between the card and system memory, but mBART-50 generates painfully slowly that way. Nothing on this page assumes offloading.

Would two GPUs run mBART-50 faster?

A second card roughly doubles the memory available but not the generation rate. With 818 cards already able to run mBART-50 alone, the case for pairing is weak.

Why does the quantisation differ between cards for mBART-50?

Each card is shown running the least-compressed copy it can hold, and mBART-50 appears at 1 different compression levels across the cards that fit it. Bigger cards get the more accurate version.

How accurate are these mBART-50 speed estimates?

They are calculated from specifications rather than measured, and each carries a range — 3,333–8,887 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 mBART-50?

The smallest card in our catalogue that holds mBART-50 is the Tesla C1080, with 4 GB of memory. It runs the model at Q8_0 using about 1.4 GB, and produces roughly 60.4 tokens per second. 818 cards in total can run it.

How fast is mBART-50 on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 5,554 tokens per second; the slowest that still runs it manages considerably less. Reading speed is around ten tokens per second, and 809 of the cards that can run mBART-50 clear that.

How much VRAM does mBART-50 need?

About 1.4 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 mBART-50 on a 8 GB GPU?

Yes. A CMP 170HX 8 GB with 8 GB runs it at Q8_0, using about 1.4 GB and generating roughly 1,035 tokens per second — a comfortable fit.

Can I run mBART-50 on a 12 GB GPU?

Yes. A GeForce RTX 3080 Ti with 12 GB runs it at Q8_0, using about 1.4 GB and generating roughly 633 tokens per second — a comfortable fit.

Can I run mBART-50 on a 16 GB GPU?

Yes. A Tesla V100 SXM2 16 GB with 16 GB runs it at Q8_0, using about 1.4 GB and generating roughly 785 tokens per second — a comfortable fit.

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