Gemma 3n TPS calculator

Open weights Google 7.9B parameters May 2025

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

582 cards that can run it

818 cards we hold specifications for

Smallest card that fits

Quadro 6000

6 GB · IQ4_XS · 16.2 tok/s

Fastest card

B200

432 tok/s · 180 GB

Which GPUs can run Gemma 3n?

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

582 cards match

Calculating

					Needs	Quantisation	Fit
432 tok/s 259–691 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	9.1 GB	Q8_0	Comfortable
432 tok/s 259–691 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	9.1 GB	Q8_0	Comfortable
345 tok/s 207–551 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	9.1 GB	Q8_0	Comfortable
345 tok/s 207–551 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	9.1 GB	Q8_0	Comfortable
276 tok/s 165–441 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	9.1 GB	Q8_0	Comfortable
264 tok/s 158–422 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	9.1 GB	Q8_0	Comfortable
264 tok/s 158–422 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	9.1 GB	Q8_0	Comfortable
253 tok/s 152–404 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	9.1 GB	Q8_0	Comfortable
224 tok/s 134–359 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	9.1 GB	Q8_0	Comfortable
224 tok/s 134–359 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	9.1 GB	Q8_0	Comfortable
224 tok/s 134–359 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	9.1 GB	Q8_0	Comfortable
213 tok/s 128–340 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	9.1 GB	Q8_0	Comfortable
181 tok/s 109–290 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	9.1 GB	Q8_0	Comfortable
181 tok/s 109–290 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	9.1 GB	Q8_0	Comfortable
181 tok/s 109–290 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	9.1 GB	Q8_0	Comfortable
181 tok/s 109–290 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	9.1 GB	Q8_0	Comfortable
181 tok/s 109–290 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	9.1 GB	Q8_0	Comfortable
144 tok/s 86–230 · low confidence	CMP 170HX 8 GB NVIDIA	8 GB	1,490 GB/s	Sep 2021	6.4 GB	Q5_K_M	Tight
138 tok/s 83–221 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	9.1 GB	Q8_0	Comfortable
138 tok/s 83–221 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	9.1 GB	Q8_0	Comfortable
122 tok/s 73–196 · low confidence	CMP 170HX 10 GB NVIDIA	10 GB	1,560 GB/s	Sep 2021	7.3 GB	Q6_K	Comfortable
115 tok/s 69–184 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	9.1 GB	Q8_0	Comfortable
113 tok/s 68–180 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	9.1 GB	Q8_0	Comfortable
110 tok/s 66–176 · low confidence	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	9.1 GB	Q8_0	Comfortable
110 tok/s 66–176 · low confidence	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	9.1 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
Organisation type: Industry
Country: United States of America
Published: 20 May 2025
Authors: Lucas Gonzalez, Rakesh Shivanna

What it does

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

Domain: Language, Multimodal, Speech, Vision
Task: Language modeling/generation, Question answering, Chat, Speech recognition (ASR), Translation, Speech-to-text, Visual question answering, Mathematical reasoning, Code generation, Character recognition (OCR)

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: 7.9B
Training data: 11,000,000,000,000 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: 5.2 × 10²³ FLOP
How it was established: Operation counting

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: google

How it is classified

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

Record confidence: Confident

Sources

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

Reference: Announcing Gemma 3n preview: powerful, efficient, mobile-first AI
Last updated: 28 November 2025

The extremes

The ten fastest GPUs that run Gemma 3n

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 3n

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

Quadro 6000

Memory needed

5.0 GB

Fastest

432 tok/s

Gemma 3n reaches a parameter count of 7.9B. 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: 582.

The entry point is Quadro 6000, with a memory capacity of 6 GB, running it at a compression of IQ4_XS and producing around 16.2 tokens per second.

The quickest result comes from B200, generating roughly 432 tokens per second on the strength of a memory bandwidth of 8,000 GB/s.

What this model is

Gemma 3n was published by Google, in the country recorded as United States of America, during May 2025. It comes out of an organisation categorised as industry.

It works in the domain of Language, Multimodal, Speech, Vision, and is recorded as performing the task of language modeling/generation, Question answering, Chat, Speech recognition (ASR), Translation, Speech-to-text, Visual question answering, Mathematical reasoning, Code generation, Character recognition (OCR).

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 google.

What decides the speed

Across every card that can run it, the middle of the range sits at 24.2 tokens per second. Clearing the ten tokens per second that roughly matches reading speed: 551 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.

How it was trained

Producing it required arithmetic totalling around 5.2 × 10²³ FLOP. 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 11,000,000,000,000 tokens of text.

Step by step

How to choose a GPU for Gemma 3n

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 Gemma 3n, needing around 5.0 GB at a compression of IQ4_XS. Capacity is the gate — a card either holds it or it does not.
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, because at long context a card that handles short questions easily can be dropped by Gemma 3n.
03

Decide how much compression you will accept

The quantisation column varies by card, because a bigger card holds a more accurate copy, reaching a compression of IQ4_XS 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

Ranking by tokens per second follows memory bandwidth rather than core counts, for Gemma 3n. It will not match a gaming ordering, because generation is bound by memory bandwidth. The card topping the list is B200, at 432 tok/s.
05

Check the fit verdict before buying

A tight fit runs, but leaves nothing spare for a longer conversation, in the case of Gemma 3n. 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

See what else that card runs

Every card name links to its own page, which runs the same calculation across the whole model catalogue. A card is usually bought for more than one model, so it is worth a look before buying for Gemma 3n.

Answers

Gemma 3n — common questions

Gemma 3n— 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 9.1 GB and generating roughly 72.3 tokens per second. The fit is comfortable.

Gemma 3n— 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.

Gemma 3n— how many parameters does it have?

It has a parameter count of 7.9B. 7.85B (Safetensors). 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.

Gemma 3n— who created it?

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

Gemma 3n— when was it released?

It was published in May 2025.

Gemma 3n— what is it used for?

It works in the domain of Language, Multimodal, Speech, Vision, and is recorded as handling the task of language modeling/generation, Question answering, Chat, Speech recognition (ASR), Translation, Speech-to-text, Visual question answering, Mathematical reasoning, Code generation, Character recognition (OCR). 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.

Gemma 3n— where can I download it?

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

Gemma 3n— how much compute was used to train it?

Training consumed around 5.2 × 10²³ FLOP. 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.

Gemma 3n— 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 0.9 GB. Every figure here assumes the whole model is resident on the card.

Gemma 3n— would two GPUs run it faster?

Two cards buy memory rather than speed, which matters only if one card cannot hold it. The number that can: 582. So a second card is rarely the answer here.

Gemma 3n— 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.

Gemma 3n— how accurate are these speed estimates?

They are calculated from specifications rather than measured, and each carries a range. One example: 259–691 tok/s on B200. The same model and card vary by thirty to fifty per cent depending on the inference software and its version.

Gemma 3n— what GPU do I need to run it?

The smallest card in our catalogue that holds it is Quadro 6000, with a memory capacity of 6 GB. It runs the model at a compression of IQ4_XS using about 5.0 GB, and produces roughly 16.2 tokens per second. The number of cards able to run it in total: 582.

Gemma 3n— how fast is it on a GPU?

It depends on the card. The quickest we calculate is B200, at about 432 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: 551.

Gemma 3n— how much VRAM does it need?

It needs about 5.0 GB at a compression of IQ4_XS, 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.

Gemma 3n— 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 Q5_K_M, using about 6.4 GB and generating roughly 144 tokens per second. The fit is tight.

Gemma 3n— 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 9.1 GB and generating roughly 49.2 tokens per second. The fit is tight.

Gemma 3n— 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 9.1 GB and generating roughly 61.0 tokens per second. The fit is comfortable.

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