Llama 3-70B TPS calculator

Open weights Meta AI 70B parameters April 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

61 cards that can run it

818 cards we hold specifications for

Smallest card that fits

A100 PCIe 40 GB

40 GB · Q3_K_M · 25.5 tok/s

Fastest card

B200

48.4 tok/s · 180 GB

Which GPUs can run Llama 3-70B?

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

61 cards match

Calculating

					Needs	Quantisation	Fit
48.4 tok/s 41–58	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	73.8 GB	Q8_0	Comfortable
48.4 tok/s 41–58	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	73.8 GB	Q8_0	Comfortable
38.7 tok/s 23–62 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	73.8 GB	Q8_0	Comfortable
38.7 tok/s 23–62 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	73.8 GB	Q8_0	Comfortable
30.9 tok/s 19–49 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	73.8 GB	Q8_0	Comfortable
29.6 tok/s 25–36	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	73.8 GB	Q8_0	Comfortable
29.6 tok/s 25–36	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	73.8 GB	Q8_0	Comfortable
29.5 tok/s 25–35	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	57.5 GB	Q6_K	Comfortable
29.5 tok/s 25–35	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	57.5 GB	Q6_K	Comfortable
28.3 tok/s 17–45 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	73.8 GB	Q8_0	Comfortable
26.1 tok/s 22–31	GRID A100B NVIDIA	48 GB	1,870 GB/s	May 2020	41.2 GB	Q4_K_M	Tight
25.5 tok/s 22–31	A100 PCIe 40 GB NVIDIA	40 GB	1,560 GB/s	Jun 2020	33.0 GB	Q3_K_M	Tight
25.5 tok/s 22–31	A100 SXM4 40 GB NVIDIA	40 GB	1,560 GB/s	May 2020	33.0 GB	Q3_K_M	Tight
25.5 tok/s 22–31	A800 PCIe 40 GB NVIDIA	40 GB	1,560 GB/s	Nov 2022	33.0 GB	Q3_K_M	Tight
25.1 tok/s 15–40 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	73.8 GB	Q8_0	Comfortable
25.1 tok/s 15–40 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	73.8 GB	Q8_0	Comfortable
25.1 tok/s 15–40 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	73.8 GB	Q8_0	Comfortable
23.8 tok/s 20–29	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	73.8 GB	Q8_0	Tight
20.3 tok/s 17–24	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	73.8 GB	Q8_0	Tight
20.3 tok/s 17–24	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	73.8 GB	Q8_0	Tight
20.3 tok/s 17–24	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	73.8 GB	Q8_0	Tight
18.7 tok/s 16–22	RTX PRO 5000 Blackwell NVIDIA	48 GB	1,340 GB/s	Mar 2025	41.2 GB	Q4_K_M	Tight
17.9 tok/s 15–22	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	57.5 GB	Q6_K	Comfortable
17.9 tok/s 15–22	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	57.5 GB	Q6_K	Comfortable
17.9 tok/s 15–22	A800 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Aug 2022	57.5 GB	Q6_K	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: 18 April 2024
Authors: Aaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Amit Sangani; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn…

What it does

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

Domain: Language
Task: Chat, Language modeling/generation, Code generation
Numerical format: BF16

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: 70B
Training data: 15,000,000,000,000 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: 7.9 × 10²⁴ 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: 6,400,000

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
Hugging Face: meta-llama

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
Why it is tracked: Significant use
Record confidence: Confident

Sources

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

Reference: Introducing Meta Llama 3: The most capable openly available LLM to date
Last updated: 18 December 2025

The extremes

The ten fastest GPUs that run Llama 3-70B

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 3-70B

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

A100 PCIe 40 GB

Memory needed

33.0 GB

Fastest

48.4 tok/s

Llama 3-70B reaches a parameter count of 70B. That puts it above consumer hardware, into the range where a card is bought for this purpose rather than repurposed for it. The number of cards we track that can hold it: 61.

At the low end it is handled by A100 PCIe 40 GB, with a memory capacity of 40 GB, running it at a compression of Q3_K_M and producing around 25.5 tokens per second.

At the other end sits B200, generating roughly 48.4 tokens per second on the strength of a memory bandwidth of 8,000 GB/s.

Background

Llama 3-70B was published by Meta AI, in the country recorded as United States of America, during April 2024. It comes out of an organisation categorised as industry.

It works in the domain of Language, and is recorded as performing the task of chat, Language modeling/generation, Code generation.

The weights are published, so it can be downloaded and run on your own hardware indefinitely, offline, with no account attached. On Hugging Face it is published under the organisation meta-llama.

Reading the throughput figures

Half the cards that hold it manage more than 17.1 tokens per second. Exceeding reading speed outright: 49 of them.

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

Its attention layout is on file, so the memory figures are computed exactly rather than approximated.

Training and provenance

The training run consumed about 7.9 × 10²⁴ 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.

It was trained on a corpus of about 15,000,000,000,000 tokens of text.

Its inclusion criterion: significant use.

Step by step

How to choose a GPU for Llama 3-70B

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

Start from what it actually needs, which is the requirement of Llama 3-70B, needing around 33.0 GB at a compression of Q3_K_M. That figure, not the headline performance of a card, is what decides whether it runs.
02

Decide how long your conversations run

The conversation occupies memory too, and grows as it goes. Set the slider to the length you expect, because at long context a card that handles short questions easily can be dropped by Llama 3-70B.
03

Set a quality floor

Each card runs the least-compressed copy it can hold, 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

Compare tokens per second, not specifications

Ranking by tokens per second follows memory bandwidth rather than core counts, for Llama 3-70B. It will not match a gaming ordering, because generation is bound by memory bandwidth. The card topping the list is B200, at 48.4 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 3-70B. 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 3-70B.

Answers

Llama 3-70B — common questions

Llama 3-70B— 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: 41–58 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 3-70B— what GPU do I need to run it?

The smallest card in our catalogue that holds it is A100 PCIe 40 GB, with a memory capacity of 40 GB. It runs the model at a compression of Q3_K_M using about 33.0 GB, and produces roughly 25.5 tokens per second. The number of cards able to run it in total: 61.

Llama 3-70B— how fast is it on a GPU?

It depends on the card. The quickest we calculate is B200, at about 48.4 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: 49.

Llama 3-70B— how much VRAM does it need?

It needs about 33.0 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.

Llama 3-70B— 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 3-70B— how many parameters does it have?

It has a parameter count of 70B. 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 3-70B— who created it?

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

Llama 3-70B— when was it released?

It was published in April 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 3-70B— what is it used for?

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

Llama 3-70B— where can I download it?

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

Llama 3-70B— how much compute was used to train it?

Training consumed around 7.9 × 10²⁴ 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 3-70B— can I run it 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 model is rarely worth using. The nearest miss we calculate falls short by 12.4 GB. Every figure here assumes the whole model is resident on the card.

Llama 3-70B— would two GPUs run it faster?

Capacity adds across cards; throughput does not. The number of cards already holding it on their own: 61. So a second card is rarely the answer here.

Llama 3-70B— 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: 4. Bigger cards get the more accurate version, and the quality floor above pins the comparison to one level.

Source

Original publication

Record last updated 18 December 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