SAM 3 TPS calculator

Open weights Meta AI 850M parameters November 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

818 cards that can run it

818 cards we hold specifications for

Smallest card that fits

Tesla C1080

4 GB · Q8_0 · 43.4 tok/s

Fastest card

B200

3,986 tok/s · 180 GB

Which GPUs can run SAM 3?

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
3,986 tok/s 2,392–6,378 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	1.6 GB	Q8_0	Comfortable
3,986 tok/s 2,392–6,378 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	1.6 GB	Q8_0	Comfortable
3,183 tok/s 1,910–5,093 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	1.6 GB	Q8_0	Comfortable
3,183 tok/s 1,910–5,093 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	1.6 GB	Q8_0	Comfortable
2,546 tok/s 1,527–4,073 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	1.6 GB	Q8_0	Comfortable
2,437 tok/s 1,462–3,898 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	1.6 GB	Q8_0	Comfortable
2,437 tok/s 1,462–3,898 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	1.6 GB	Q8_0	Comfortable
2,332 tok/s 1,399–3,731 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	1.6 GB	Q8_0	Comfortable
2,070 tok/s 1,242–3,311 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	1.6 GB	Q8_0	Comfortable
2,070 tok/s 1,242–3,311 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	1.6 GB	Q8_0	Comfortable
2,070 tok/s 1,242–3,311 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	1.6 GB	Q8_0	Comfortable
1,963 tok/s 1,178–3,141 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	1.6 GB	Q8_0	Comfortable
1,674 tok/s 1,005–2,679 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	1.6 GB	Q8_0	Comfortable
1,674 tok/s 1,005–2,679 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	1.6 GB	Q8_0	Comfortable
1,674 tok/s 1,005–2,679 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	1.6 GB	Q8_0	Comfortable
1,674 tok/s 1,005–2,679 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	1.6 GB	Q8_0	Comfortable
1,674 tok/s 1,005–2,679 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	1.6 GB	Q8_0	Comfortable
1,275 tok/s 765–2,040 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	1.6 GB	Q8_0	Comfortable
1,275 tok/s 765–2,040 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	1.6 GB	Q8_0	Comfortable
1,062 tok/s 637–1,700 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	1.6 GB	Q8_0	Comfortable
1,040 tok/s 624–1,663 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	1.6 GB	Q8_0	Comfortable
1,016 tok/s 610–1,626 · low confidence	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	1.6 GB	Q8_0	Comfortable
1,016 tok/s 610–1,626 · low confidence	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	1.6 GB	Q8_0	Comfortable
1,016 tok/s 610–1,626 · low confidence	A800 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Aug 2022	1.6 GB	Q8_0	Comfortable
1,016 tok/s 610–1,626 · low confidence	H100 CNX NVIDIA	80 GB	2,040 GB/s	Mar 2023	1.6 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: Meta AI
Organisation type: Industry
Country: United States of America
Published: 20 November 2025
Authors: Nicolas Carion, Laura Gustafson, Yuan-Ting Hu, Shoubhik Debnath, Ronghang Hu, Didac Suris, Chaitanya Ryali, Kalyan Vasudev Alwala, Haitham Khedr, Andrew Huang, Jie Lei, Tengyu Ma, Baishan Guo, Arpit Kalla, Markus Marks, Joseph Greer, Meng Wang, Peize Sun, Roman Rädle, Triantafyllos Afouras, Effrosyni Mavroudi, Katherine Xu, Tsung-Han Wu, Yu Zhou, Liliane Momeni, Rishi Hazra, Shuangrui Ding, Sagar …

What it does

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

Domain: Vision
Task: Image segmentation

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: 850M
Training data: 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: 2 × 10²³ FLOP

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)

How it is classified

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

Record confidence: Likely

Sources

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

Reference: SAM 3: Segment Anything with Concepts
Last updated: 8 April 2026

The extremes

The ten fastest GPUs that run SAM 3

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 SAM 3

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

1.6 GB

Fastest

3,986 tok/s

SAM 3 is small enough at 850M parameters that hardware is rarely the obstacle — 818 of the cards we track can run it, including cards several years old.

The smallest card that holds it is the Tesla C1080 with 4 GB, running it at Q8_0 and producing around 43.4 tokens per second.

Top of the range is the B200, at roughly 3,986 tokens per second thanks to 8,000 GB/s of bandwidth.

Where it came from

SAM 3 was published by Meta AI, in United States of America, in November 2025. It comes out of industry.

It works in Vision, and is recorded as doing image segmentation.

The weights being open is what puts this page in the calculator rather than only in the catalogue: it is a model you can actually hold.

Understanding the speeds

Across every card that can run it, the middle of the range is about 111.9 tokens per second, and 809 of them clear the ten tokens per second that roughly matches reading speed.

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

Its internal architecture is not on file, so memory is approximated from the parameter count and marked accordingly. Expect the real figure to differ, more so at long context.

Training and provenance

Training it took roughly 2 × 10²³ FLOP of computation — a measure of what producing the model cost, not of how fast it answers.

Step by step

How to choose a GPU for SAM 3

The table above has already assessed every card we hold specifications for against this model. Getting to your answer takes six steps.

01

Check what it needs before anything else

Look at what SAM 3 actually needs — around 1.6 GB at Q8_0. No amount of processing power compensates for a card that cannot hold it.
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: at long context SAM 3 can slip off a card that handles short questions easily.
03

Set a quality floor

Each card runs the least-compressed copy it can hold — Q8_0 on the smallest card that fits. Setting a floor drops the cards that only manage SAM 3 by squeezing it further than you would want.
04

Sort by speed

The speed ordering for SAM 3 is effectively an ordering by memory bandwidth, which is why the B200 tops it at 3,986 tok/s.
05

Read the fit column last

The fit column separates cards that just manage SAM 3 from those with room to spare. Buy for the second if the context might grow.
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 SAM 3 is settled.

Answers

SAM 3 — common questions

Would two GPUs run SAM 3 faster?

Two cards buy memory rather than speed. That matters for SAM 3 only if one card cannot hold it — 818 can, so a second adds little.

Why does the quantisation differ between cards for SAM 3?

Each card is shown running the least-compressed copy it can hold, and SAM 3 appears at 1 different compression levels across the cards that fit it. Bigger cards get the more accurate version.

How accurate are these SAM 3 speed estimates?

These are estimates with real error bars. The fastest result here, 2,392–6,378 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 SAM 3?

The smallest card in our catalogue that holds SAM 3 is the Tesla C1080, with 4 GB of memory. It runs the model at Q8_0 using about 1.6 GB, and produces roughly 43.4 tokens per second. 818 cards in total can run it.

How fast is SAM 3 on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 3,986 tokens per second; the slowest that still runs it manages considerably less. Reading speed is around ten tokens per second, and 809 of the cards that can run SAM 3 clear that.

How much VRAM does SAM 3 need?

About 1.6 GB at Q8_0 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 SAM 3 on a 8 GB GPU?

Yes. A CMP 170HX 8 GB with 8 GB runs it at Q8_0, using about 1.6 GB and generating roughly 742 tokens per second — a comfortable fit.

Can I run SAM 3 on a 12 GB GPU?

Yes. A GeForce RTX 3080 Ti with 12 GB runs it at Q8_0, using about 1.6 GB and generating roughly 455 tokens per second — a comfortable fit.

Can I run SAM 3 on a 16 GB GPU?

Yes. A Tesla V100 SXM2 16 GB with 16 GB runs it at Q8_0, using about 1.6 GB and generating roughly 563 tokens per second — a comfortable fit.

Can I run SAM 3 on a 24 GB GPU?

Yes. A GeForce RTX 5090 D V2 with 24 GB runs it at Q8_0, using about 1.6 GB and generating roughly 668 tokens per second — a comfortable fit.

Is SAM 3 open source?

Its weights are published, so SAM 3 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 SAM 3 have?

SAM 3 has 850M parameters. "SAM 3 has ∼850M parameters, distributed as follows: ∼450M and ∼300M for the vision and text encoders (Bolya et al., 2025), and ∼100M for the detector and tracker components". 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 SAM 3?

SAM 3 was published by Meta AI, based in United States of America, categorised as industry.

When was SAM 3 released?

SAM 3 was published in November 2025.

What is SAM 3 used for?

SAM 3 works in Vision, and is recorded as handling image segmentation. These are the areas it was designed around; they describe intent rather than a hard boundary.

Where can I download SAM 3?

The weights for SAM 3 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 SAM 3?

Around 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.

Can I run SAM 3 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. Our figures for SAM 3 assume it is fully resident.

Source

Original publication

Record last updated 8 April 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