Llama 3.2 3B TPS calculator

Open weights Meta AI 3.2B parameters September 2024

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 · Q4_K_M · 26.5 tok/s

Fastest card

B200

1,056 tok/s · 180 GB

Which GPUs can run Llama 3.2 3B?

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
1,056 tok/s 897–1,267	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	4.8 GB	Q8_0	Comfortable
1,056 tok/s 897–1,267	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	4.8 GB	Q8_0	Comfortable
843 tok/s 506–1,349 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	4.8 GB	Q8_0	Comfortable
843 tok/s 506–1,349 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	4.8 GB	Q8_0	Comfortable
674 tok/s 404–1,079 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	4.8 GB	Q8_0	Comfortable
645 tok/s 548–774	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	4.8 GB	Q8_0	Comfortable
645 tok/s 548–774	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	4.8 GB	Q8_0	Comfortable
617 tok/s 370–988 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	4.8 GB	Q8_0	Comfortable
548 tok/s 329–877 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	4.8 GB	Q8_0	Comfortable
548 tok/s 329–877 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	4.8 GB	Q8_0	Comfortable
548 tok/s 329–877 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	4.8 GB	Q8_0	Comfortable
520 tok/s 442–624	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	4.8 GB	Q8_0	Comfortable
443 tok/s 377–532	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	4.8 GB	Q8_0	Comfortable
443 tok/s 377–532	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	4.8 GB	Q8_0	Comfortable
443 tok/s 377–532	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	4.8 GB	Q8_0	Comfortable
443 tok/s 377–532	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	4.8 GB	Q8_0	Comfortable
443 tok/s 377–532	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	4.8 GB	Q8_0	Comfortable
338 tok/s 203–540 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	4.8 GB	Q8_0	Comfortable
338 tok/s 203–540 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	4.8 GB	Q8_0	Comfortable
281 tok/s 169–450 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	4.8 GB	Q8_0	Comfortable
275 tok/s 165–440 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	4.8 GB	Q8_0	Comfortable
269 tok/s 229–323	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	4.8 GB	Q8_0	Comfortable
269 tok/s 229–323	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	4.8 GB	Q8_0	Comfortable
269 tok/s 229–323	A800 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Aug 2022	4.8 GB	Q8_0	Comfortable
269 tok/s 229–323	H100 CNX NVIDIA	80 GB	2,040 GB/s	Mar 2023	4.8 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: Meta AI
Organisation type: Industry
Country: United States of America
Published: 24 September 2024

What it does

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

Domain: Language
Task: Language modeling/generation, Text summarization, Question answering, Quantitative reasoning, 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: 3.2B
Training data: 9,000,000,000,000 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: 1.7 × 10²³ FLOP
How it was established: Operation counting,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 H100 SXM5 80GB
Chip-hours: 460,000

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 (restricted use)
Training code: Unreleased

How it is classified

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

Likely above 10²³ FLOP: Yes
Record confidence: Confident

Sources

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

Reference: Llama 3.2: Revolutionizing edge AI and vision with open, customizable models
Last updated: 28 November 2025

The extremes

The ten fastest GPUs that run Llama 3.2 3B

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 Llama 3.2 3B

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

What it takes to run this model

Minimum card

Tesla C1080

Memory needed

3.3 GB

Fastest

1,056 tok/s

Llama 3.2 3B reaches a parameter count of 3.2B. 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: 818.

The smallest card that holds it is Tesla C1080, with a memory capacity of 4 GB, running it at a compression of Q4_K_M and producing around 26.5 tokens per second.

The quickest result comes from B200, generating roughly 1,056 tokens per second on the strength of a memory bandwidth of 8,000 GB/s.

What this model is

Llama 3.2 3B was published by Meta AI, in the country recorded as United States of America, during September 2024. The category the publisher falls under is industry.

It works in the domain of Language, and is recorded as performing the task of language modeling/generation, Text summarization, Question answering, Quantitative reasoning, Translation.

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

Across every card that can run it, the middle of the range sits at 39.0 tokens per second. Exceeding reading speed outright: 785 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.

Because the architecture is recorded, the memory column is derived rather than estimated.

Training and provenance

Producing it required arithmetic totalling around 1.7 × 10²³ FLOP, on hardware recorded as NVIDIA H100 SXM5 80GB. That figure measures what producing the model cost, and has no bearing on how fast it answers.

The training set ran to roughly 9,000,000,000,000 tokens of text.

Step by step

How to choose a GPU for Llama 3.2 3B

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

The table lists every card able to hold Llama 3.2 3B, needing around 3.3 GB at a compression of Q4_K_M. Capacity is the gate — a card either holds it or it does not.
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 a card that seemed fine stops fitting Llama 3.2 3B.
03

Choose how far you will compress it

Each card runs the least-compressed copy it can hold, reaching a compression of Q4_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

Sort by speed

Ranking by tokens per second follows memory bandwidth rather than core counts, for Llama 3.2 3B. It will not match a gaming ordering, because generation is bound by memory bandwidth. The card topping the list is B200, at 1,056 tok/s.
05

Look at the headroom, not just the fit

A tight fit runs, but leaves nothing spare for a longer conversation, in the case of Llama 3.2 3B. 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 Llama 3.2 3B.

Answers

Llama 3.2 3B — common questions

Llama 3.2 3B— how fast is it on a GPU?

It depends on the card. The quickest we calculate is B200, at about 1,056 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: 785.

Llama 3.2 3B— how much VRAM does it need?

It needs about 3.3 GB at a compression of Q4_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.

Llama 3.2 3B— can I run it on a GPU holding 8 GB?

Yes. The card CMP 170HX 8 GB, holding 8 GB, runs it at a compression of Q8_0, using about 4.8 GB and generating roughly 197 tokens per second. The fit is comfortable.

Llama 3.2 3B— 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 Q8_0, using about 4.8 GB and generating roughly 120 tokens per second. The fit is comfortable.

Llama 3.2 3B— 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 Q8_0, using about 4.8 GB and generating roughly 149 tokens per second. The fit is comfortable.

Llama 3.2 3B— 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 4.8 GB and generating roughly 177 tokens per second. The fit is comfortable.

Llama 3.2 3B— 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.

Llama 3.2 3B— how many parameters does it have?

It has a parameter count of 3.2B. https://huggingface.co/meta-llama/Llama-3.2-1B. 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.

Llama 3.2 3B— who created it?

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

Llama 3.2 3B— when was it released?

It was published in September 2024. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.

Llama 3.2 3B— what is it used for?

It works in the domain of Language, and is recorded as handling the task of language modeling/generation, Text summarization, Question answering, Quantitative reasoning, Translation. A model can carry several of each, so these are the areas it was built for rather than a limit on what it will attempt.

Llama 3.2 3B— 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.

Llama 3.2 3B— how much compute was used to train it?

Training consumed around 1.7 × 10²³ FLOP, on hardware recorded as NVIDIA H100 SXM5 80GB. 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.

Llama 3.2 3B— can I run it if it does not fit in my GPU?

Only by offloading, which is usually a false economy: the part held in system memory drags the whole thing down. Every figure here assumes the whole model is resident on the card.

Llama 3.2 3B— 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: 818. So a second card is rarely the answer here.

Llama 3.2 3B— why does the quantisation differ between cards?

Because capacity varies, so does how hard it has to be squeezed. The number of distinct levels in the table above: 3. Bigger cards get the more accurate version, and the quality floor above pins the comparison to one level.

Llama 3.2 3B— how accurate are these speed estimates?

These are estimates with real error bars, and any of them could reasonably land anywhere in its published range depending on which runtime you use. The fastest result here: 897–1,267 tok/s on B200. The same model and card vary by thirty to fifty per cent depending on the inference software and its version.

Llama 3.2 3B— what GPU do I need to run it?

The smallest card in our catalogue that holds it is Tesla C1080, with a memory capacity of 4 GB. It runs the model at a compression of Q4_K_M using about 3.3 GB, and produces roughly 26.5 tokens per second. The number of cards able to run it in total: 818.

Source

Original publication

Record last updated 28 November 2025

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