Gemma 7B TPS calculator

Open weights Google DeepMind 8.5B parameters February 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

306 cards that can run it

818 cards we hold specifications for

Smallest card that fits

P102-101

10 GB · Q4_K_M · 31.2 tok/s

Fastest card

B200

397 tok/s · 180 GB

Which GPUs can run Gemma 7B?

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

306 cards match

Calculating

					Needs	Quantisation	Fit
397 tok/s 337–476	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	12.8 GB	Q8_0	Comfortable
397 tok/s 337–476	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	12.8 GB	Q8_0	Comfortable
317 tok/s 190–507 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	12.8 GB	Q8_0	Comfortable
317 tok/s 190–507 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	12.8 GB	Q8_0	Comfortable
253 tok/s 152–405 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	12.8 GB	Q8_0	Comfortable
243 tok/s 206–291	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	12.8 GB	Q8_0	Comfortable
243 tok/s 206–291	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	12.8 GB	Q8_0	Comfortable
232 tok/s 139–371 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	12.8 GB	Q8_0	Comfortable
206 tok/s 124–330 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	12.8 GB	Q8_0	Comfortable
206 tok/s 124–330 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	12.8 GB	Q8_0	Comfortable
206 tok/s 124–330 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	12.8 GB	Q8_0	Comfortable
195 tok/s 166–235	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	12.8 GB	Q8_0	Comfortable
179 tok/s 152–214	CMP 170HX 10 GB NVIDIA	10 GB	1,560 GB/s	Sep 2021	8.8 GB	Q4_K_M	Tight
167 tok/s 142–200	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	12.8 GB	Q8_0	Comfortable
167 tok/s 142–200	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	12.8 GB	Q8_0	Comfortable
167 tok/s 142–200	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	12.8 GB	Q8_0	Comfortable
167 tok/s 142–200	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	12.8 GB	Q8_0	Comfortable
167 tok/s 142–200	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	12.8 GB	Q8_0	Comfortable
127 tok/s 76–203 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	12.8 GB	Q8_0	Comfortable
127 tok/s 76–203 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	12.8 GB	Q8_0	Comfortable
106 tok/s 63–169 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	12.8 GB	Q8_0	Comfortable
104 tok/s 62–166 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	12.8 GB	Q8_0	Comfortable
101 tok/s 86–121	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	12.8 GB	Q8_0	Comfortable
101 tok/s 86–121	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	12.8 GB	Q8_0	Comfortable
101 tok/s 86–121	A800 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Aug 2022	12.8 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: Google DeepMind
Organisation type: Industry
Country: United States of America
Published: 21 February 2024
Authors: Gemma Team, Google DeepMind

What it does

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

Domain: Language
Task: Language modeling/generation, Chat, Code generation, Question answering, Quantitative reasoning

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: 8.5B
Training data: 6,000,000,000,000 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: 3.1 × 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: Google TPU v5e
Chips used: 4,096
Power draw: 1.8 MW

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: Gemma: Open Models Based on Gemini Research and Technology
Last updated: 28 November 2025

The extremes

The ten fastest GPUs that run Gemma 7B

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 Gemma 7B

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 you need to run it

Minimum card

P102-101

Memory needed

8.8 GB

Fastest

397 tok/s

Gemma 7B is small enough at 8.5B parameters that hardware is rarely the obstacle — 306 of the cards we track can run it, including cards several years old.

The least hardware that works is a P102-101. Its 10 GB is enough at Q4_K_M compression, giving roughly 31.2 tokens per second.

At the other end, a B200 generates roughly 397 tokens per second on it, on the strength of 8,000 GB/s of memory bandwidth.

What this model is

Gemma 7B was published by Google DeepMind, in United States of America, in February 2024. It comes out of industry.

It works in Language, and is recorded as doing language modeling/generation, Chat, Code generation, Question answering, Quantitative reasoning.

Because its weights were released, nothing about running it depends on a provider staying available — it is yours once downloaded.

What decides the speed

Half the cards that hold it manage more than 30.4 tokens per second, and 282 exceed reading speed outright.

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.

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

Training and provenance

Training it took roughly 3.1 × 10²³ FLOP of computation, on Google TPU v5e — a measure of what producing the model cost, not of how fast it answers.

Around 6,000,000,000,000 tokens went into training it.

Step by step

How to choose a GPU for Gemma 7B

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

Look at what Gemma 7B actually needs — around 8.8 GB at Q4_K_M. No amount of processing power compensates for a card that cannot hold it.
02

Match the context to your actual use

The conversation occupies memory too, and grows as it goes. Set the slider to the length you expect: at long context Gemma 7B can slip off a card that handles short questions easily.
03

Set a quality floor

Compression is what makes Gemma 7B fit smaller cards, at some cost in accuracy — Q4_K_M on the smallest card that fits. A minimum quality removes the ones that go too far.
04

Rank by throughput rather than spec sheet

Sort by speed to see how cards rank for Gemma 7B. It will not match a gaming ordering — generation is bound by memory bandwidth, which is why the B200 tops it at 397 tok/s.
05

Check the fit verdict before buying

The fit column separates cards that just manage Gemma 7B from those with room to spare. Buy for the second if the context might grow.
06

See what else that card runs

Every card name links to its own page, which runs the same calculation across the whole model catalogue. Worth a look before buying for Gemma 7B alone — a card is usually bought for more than one model.

Answers

Gemma 7B — common questions

When was Gemma 7B released?

Gemma 7B was published in February 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.

What is Gemma 7B used for?

Gemma 7B works in Language, and is recorded as handling language modeling/generation, Chat, Code generation, Question answering, Quantitative reasoning. 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.

Where can I download Gemma 7B?

The weights for Gemma 7B 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 Gemma 7B?

Around 3.1 × 10²³ FLOP, on Google TPU v5e. 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 Gemma 7B if it does not fit in my GPU?

Only by offloading, which is usually a false economy: the part in system memory drags the whole thing down — the nearest miss we calculate is short by 1.6 GB. Our figures for Gemma 7B assume it is fully resident.

Would two GPUs run Gemma 7B faster?

Capacity adds across cards; throughput does not. Since 306 of the cards we track already hold Gemma 7B on their own, a second card is rarely the answer here.

Why does the quantisation differ between cards for Gemma 7B?

Because capacity varies, so does how hard Gemma 7B has to be squeezed — 3 distinct levels appear in the table above. Set a minimum quality to compare at one.

How accurate are these Gemma 7B speed estimates?

These are estimates with real error bars. The fastest result here, 337–476 tok/s on the B200, could reasonably land anywhere in its published range depending on which runtime you use.

What GPU do I need to run Gemma 7B?

The smallest card in our catalogue that holds Gemma 7B is the P102-101, with 10 GB of memory. It runs the model at Q4_K_M using about 8.8 GB, and produces roughly 31.2 tokens per second. 306 cards in total can run it.

How fast is Gemma 7B on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 397 tokens per second; the slowest that still runs it manages considerably less. Reading speed is around ten tokens per second, and 282 of the cards that can run Gemma 7B clear that.

How much VRAM does Gemma 7B need?

About 8.8 GB at Q4_K_M 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 Gemma 7B on a 12 GB GPU?

Yes. A GeForce RTX 3080 Ti with 12 GB runs it at Q5_K_M, using about 9.8 GB and generating roughly 80.8 tokens per second — a tight fit.

Can I run Gemma 7B on a 16 GB GPU?

Yes. A Tesla V100 SXM2 16 GB with 16 GB runs it at Q8_0, using about 12.8 GB and generating roughly 56.1 tokens per second — a tight fit.

Can I run Gemma 7B on a 24 GB GPU?

Yes. A GeForce RTX 5090 D V2 with 24 GB runs it at Q8_0, using about 12.8 GB and generating roughly 66.5 tokens per second — a comfortable fit.

Is Gemma 7B open source?

Its weights are published, so Gemma 7B 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 Gemma 7B have?

Gemma 7B has 8.5B parameters. Table 2, sum of embedding and non-embedding parameters. 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 Gemma 7B?

Gemma 7B was published by Google DeepMind, based in United States of America, categorised as industry.

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