Flan UL2 TPS calculator

Open weights Google Brain 19.5B parameters March 2023

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

293 cards that can run it

818 cards we hold specifications for

Smallest card that fits

Quadro K6000

12 GB · Q3_K_M · 14.4 tok/s

Fastest card

B200

174 tok/s · 180 GB

Which GPUs can run Flan UL2?

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

293 cards match

Calculating

					Needs	Quantisation	Fit
174 tok/s 104–278 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	21.6 GB	Q8_0	Comfortable
174 tok/s 104–278 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	21.6 GB	Q8_0	Comfortable
139 tok/s 83–222 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	21.6 GB	Q8_0	Comfortable
139 tok/s 83–222 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	21.6 GB	Q8_0	Comfortable
111 tok/s 67–178 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	21.6 GB	Q8_0	Comfortable
106 tok/s 64–170 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	21.6 GB	Q8_0	Comfortable
106 tok/s 64–170 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	21.6 GB	Q8_0	Comfortable
102 tok/s 61–163 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	21.6 GB	Q8_0	Comfortable
90.2 tok/s 54–144 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	21.6 GB	Q8_0	Comfortable
90.2 tok/s 54–144 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	21.6 GB	Q8_0	Comfortable
90.2 tok/s 54–144 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	21.6 GB	Q8_0	Comfortable
85.6 tok/s 51–137 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	21.6 GB	Q8_0	Comfortable
73.0 tok/s 44–117 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	21.6 GB	Q8_0	Comfortable
73.0 tok/s 44–117 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	21.6 GB	Q8_0	Comfortable
73.0 tok/s 44–117 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	21.6 GB	Q8_0	Comfortable
73.0 tok/s 44–117 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	21.6 GB	Q8_0	Comfortable
73.0 tok/s 44–117 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	21.6 GB	Q8_0	Comfortable
56.7 tok/s 34–91 · low confidence	Tesla V100 SXM2 16 GB NVIDIA	16 GB	1,130 GB/s	Nov 2019	12.5 GB	Q4_K_M	Tight
55.6 tok/s 33–89 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	21.6 GB	Q8_0	Comfortable
55.6 tok/s 33–89 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	21.6 GB	Q8_0	Comfortable
53.5 tok/s 32–86 · low confidence	GeForce RTX 3080 12 GB NVIDIA	12 GB	912 GB/s	Jan 2022	10.2 GB	Q3_K_M	Tight
53.5 tok/s 32–86 · low confidence	GeForce RTX 3080 Ti NVIDIA	12 GB	912 GB/s	May 2021	10.2 GB	Q3_K_M	Tight
48.1 tok/s 29–77 · low confidence	GeForce RTX 5080 NVIDIA	16 GB	960 GB/s	Jan 2025	12.5 GB	Q4_K_M	Tight
46.3 tok/s 28–74 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	21.6 GB	Q8_0	Comfortable
45.3 tok/s 27–73 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	21.6 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: Google Brain
Organisation type: Industry
Country: United States of America
Published: 3 March 2023
Authors: Yi Tay, Mostafa Dehghani, Vinh Q. Tran, Xavier Garcia, Jason Wei, Xuezhi Wang, Hyung Won Chung, Siamak Shakeri, Dara Bahri, Tal Schuster, Huaixiu Steven Zheng, Denny Zhou, Neil Houlsby, Donald Metzler

What it does

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

Domain: Language
Task: Language generation
Base model: UL2

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: 19.5B
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

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: 387

Sources

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

Reference: A New Open Source Flan 20B with UL2
Last updated: 25 May 2026

The extremes

The ten fastest GPUs that run Flan UL2

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 Flan UL2

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

Quadro K6000

Memory needed

10.2 GB

Fastest

174 tok/s

Flan UL2 reaches a parameter count of 19.5B. That is small enough that hardware is rarely the obstacle, including on cards several years old. The number of cards we track that can run it: 293.

At the low end it is handled by Quadro K6000, with a memory capacity of 12 GB, running it at a compression of Q3_K_M and producing around 14.4 tokens per second.

At the other end sits B200, generating roughly 174 tokens per second on the strength of a memory bandwidth of 8,000 GB/s.

What this model is

Flan UL2 was published by Google Brain, in the country recorded as United States of America, during March 2023. The publishing organisation is categorised as industry.

It works in the domain of Language, and is recorded as performing the task of language generation.

It builds on UL2. That is the usual way a specialised model is produced.

The weights are published, so it can be downloaded and run on your own hardware indefinitely, offline, with no account attached.

What decides the speed

Half the cards that hold it manage more than 19.8 tokens per second. Clearing the ten tokens per second that roughly matches reading speed: 249 of them.

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.

Step by step

How to choose a GPU for Flan UL2

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 Flan UL2, needing around 10.2 GB at a compression of Q3_K_M. No amount of processing power compensates for a card that cannot hold it.
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 Flan UL2.
03

Decide how much compression you will accept

Compression is what makes a model fit smaller cards, at some cost in accuracy, reaching a compression of Q3_K_M on the smallest card that fits. Setting a minimum quality drops the cards that only manage it by squeezing further than you would want, and holds the comparison at one level.
04

Rank by throughput rather than spec sheet

The speed ordering is effectively an ordering by memory bandwidth, for Flan UL2. It will not match a gaming ordering, because generation is bound by memory bandwidth. The card topping the list is B200, at 174 tok/s.
05

Check the fit verdict before buying

Tight means it loads and works with no room to raise the context later, in the case of Flan UL2. Comfortable means you can grow the context later. That difference matters more than a few tokens per second, so buy for comfortable if you expect to.
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 you have settled on Flan UL2.

Answers

Flan UL2 — common questions

Flan UL2— can I run it if it does not fit in my GPU?

It can be split between the card and system memory, but it generates painfully slowly that way. The nearest miss we calculate falls short by 2.6 GB. Every figure here assumes the whole model is resident on the card.

Flan UL2— would two GPUs run it faster?

A second card roughly doubles the memory available but not the generation rate. The number already able to run it alone: 293. So a second card is rarely the answer here.

Flan UL2— why does the quantisation differ between cards?

A larger card holds a more accurate copy. The number of compression levels used across the cards that run it: 4. Bigger cards get the more accurate version, and the quality floor above pins the comparison to one level.

Flan UL2— how accurate are these speed estimates?

They are calculated from specifications rather than measured, and each carries a range. One example: 104–278 tok/s on B200. The same model and card vary by thirty to fifty per cent depending on the inference software and its version.

Flan UL2— what GPU do I need to run it?

The smallest card in our catalogue that holds it is Quadro K6000, with a memory capacity of 12 GB. It runs the model at a compression of Q3_K_M using about 10.2 GB, and produces roughly 14.4 tokens per second. The number of cards able to run it in total: 293.

Flan UL2— how fast is it on a GPU?

It depends on the card. The quickest we calculate is B200, at about 174 tokens per second; the slowest that still runs it manages considerably less. Reading speed is around ten tokens per second, and the number of cards clearing that: 249.

Flan UL2— how much VRAM does it need?

It needs about 10.2 GB at a compression of Q3_K_M, 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.

Flan UL2— can I run it on a GPU holding 12 GB?

Yes. The card GeForce RTX 3080 Ti, holding 12 GB, runs it at a compression of Q3_K_M, using about 10.2 GB and generating roughly 53.5 tokens per second. The fit is tight.

Flan UL2— can I run it on a GPU holding 16 GB?

Yes. The card Tesla V100 SXM2 16 GB, holding 16 GB, runs it at a compression of Q4_K_M, using about 12.5 GB and generating roughly 56.7 tokens per second. The fit is tight.

Flan UL2— can I run it on a GPU holding 24 GB?

Yes. The card GeForce RTX 5090 D V2, holding 24 GB, runs it at a compression of Q8_0, using about 21.6 GB and generating roughly 29.1 tokens per second. The fit is tight.

Flan UL2— is it open source?

Its weights are published, so it 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.

Flan UL2— how many parameters does it have?

It has a parameter count of 19.5B. 19.5B per https://www.yitay.net/blog/flan-ul2-20b. 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.

Flan UL2— who created it?

It was published by Google Brain, based in United States of America, an organisation categorised as industry.

Flan UL2— when was it released?

It was published in March 2023. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.

Flan UL2— what is it used for?

It works in the domain of Language, and is recorded as handling the task of language generation. Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.

Flan UL2— where can I download it?

The weights are published, though we do not hold a repository link for it. This site calculates hardware requirements rather than hosting model files.

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