MedSigLIP TPS calculator

Open weights Google 800M parameters July 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 · 46.1 tok/s

Fastest card

B200

4,235 tok/s · 180 GB

Which GPUs can run MedSigLIP?

Set the inputs, read the answer

A longer conversation needs more memory, which can push this model off smaller cards.

Hides cards that would only fit the model by compressing it below this point.

818 cards match

Calculating
Needs Quantisation Fit
4,235 tok/s

2,541–6,776 · low confidence

B200 NVIDIA 180 GB 8,000 GB/s Jan 2024 1.6 GB Q8_0 Comfortable
4,235 tok/s

2,541–6,776 · low confidence

B300 NVIDIA 288 GB 8,000 GB/s Sep 2025 1.6 GB Q8_0 Comfortable
3,382 tok/s

2,029–5,411 · low confidence

Radeon Instinct MI350X AMD 288 GB 8,190 GB/s Jan 2025 1.6 GB Q8_0 Comfortable
3,382 tok/s

2,029–5,411 · low confidence

Radeon Instinct MI355X AMD 288 GB 8,190 GB/s Jan 2025 1.6 GB Q8_0 Comfortable
2,705 tok/s

1,623–4,328 · low confidence

Radeon Instinct MI300 AMD 128 GB 6,550 GB/s Jan 2023 1.6 GB Q8_0 Comfortable
2,589 tok/s

1,553–4,142 · low confidence

H200 NVL NVIDIA 141 GB 4,890 GB/s Nov 2024 1.6 GB Q8_0 Comfortable
2,589 tok/s

1,553–4,142 · low confidence

H200 SXM 141 GB NVIDIA 141 GB 4,890 GB/s Nov 2024 1.6 GB Q8_0 Comfortable
2,478 tok/s

1,487–3,964 · low confidence

Radeon Instinct MI325X AMD 256 GB 6,000 GB/s Oct 2024 1.6 GB Q8_0 Comfortable
2,199 tok/s

1,319–3,518 · low confidence

Radeon Instinct MI300A AMD 128 GB 5,325 GB/s Dec 2023 1.6 GB Q8_0 Comfortable
2,199 tok/s

1,319–3,518 · low confidence

Radeon Instinct MI300X AMD 192 GB 5,325 GB/s Dec 2023 1.6 GB Q8_0 Comfortable
2,199 tok/s

1,319–3,518 · low confidence

Radeon Instinct MI308X AMD 192 GB 5,325 GB/s Dec 2023 1.6 GB Q8_0 Comfortable
2,086 tok/s

1,252–3,337 · low confidence

H100 NVL 94 GB NVIDIA 94 GB 3,940 GB/s Mar 2023 1.6 GB Q8_0 Comfortable
1,779 tok/s

1,067–2,846 · low confidence

H100 PCIe 96 GB NVIDIA 96 GB 3,360 GB/s Mar 2023 1.6 GB Q8_0 Comfortable
1,779 tok/s

1,067–2,846 · low confidence

H100 SXM5 80 GB NVIDIA 80 GB 3,360 GB/s Oct 2022 1.6 GB Q8_0 Comfortable
1,779 tok/s

1,067–2,846 · low confidence

H100 SXM5 94 GB NVIDIA 94 GB 3,360 GB/s Mar 2023 1.6 GB Q8_0 Comfortable
1,779 tok/s

1,067–2,846 · low confidence

H100 SXM5 96 GB NVIDIA 96 GB 3,360 GB/s Mar 2023 1.6 GB Q8_0 Comfortable
1,779 tok/s

1,067–2,846 · low confidence

H800 SXM5 NVIDIA 80 GB 3,360 GB/s Mar 2023 1.6 GB Q8_0 Comfortable
1,354 tok/s

813–2,167 · low confidence

Radeon Instinct MI250 AMD 128 GB 3,280 GB/s Nov 2021 1.6 GB Q8_0 Comfortable
1,354 tok/s

813–2,167 · low confidence

Radeon Instinct MI250X AMD 128 GB 3,280 GB/s Nov 2021 1.6 GB Q8_0 Comfortable
1,129 tok/s

677–1,806 · low confidence

Data Center GPU Max 1550 Intel 128 GB 3,280 GB/s Jan 2023 1.6 GB Q8_0 Comfortable
1,105 tok/s

663–1,767 · low confidence

Data Center GPU Max Subsystem Intel 128 GB 3,210 GB/s Jan 2023 1.6 GB Q8_0 Comfortable
1,080 tok/s

648–1,728 · low confidence

A100 SXM4 80 GB NVIDIA 80 GB 2,040 GB/s Nov 2020 1.6 GB Q8_0 Comfortable
1,080 tok/s

648–1,728 · low confidence

A100X NVIDIA 80 GB 2,040 GB/s Jun 2021 1.6 GB Q8_0 Comfortable
1,080 tok/s

648–1,728 · low confidence

A800 SXM4 80 GB NVIDIA 80 GB 2,040 GB/s Aug 2022 1.6 GB Q8_0 Comfortable
1,080 tok/s

648–1,728 · 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
Google
Organisation type
Industry
Country
United States of America
Published
9 July 2025

What it does

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

Domain
Vision, Medicine
Task
Image embedding, Image segmentation, Image classification

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
800M

"MedSigLIP contains a 400M parameter vision encoder and 400M parameter text encoder,"

Training data
tokens

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

The use of MedSigLIP is governed by the Health AI Developer Foundations terms of use https://huggingface.co/google/medsiglip-448

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
MedSigLIP model card
Last updated
28 November 2025

The extremes

The ten fastest GPUs that run MedSigLIP

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.

  1. 01 B300 288 GB · 8,000 GB/s · Q8_0 4,235 tok/s
  2. 02 B200 180 GB · 8,000 GB/s · Q8_0 4,235 tok/s
  3. 03 Radeon Instinct MI350X 288 GB · 8,190 GB/s · Q8_0 3,382 tok/s
  4. 04 Radeon Instinct MI355X 288 GB · 8,190 GB/s · Q8_0 3,382 tok/s
  5. 05 Radeon Instinct MI300 128 GB · 6,550 GB/s · Q8_0 2,705 tok/s
  6. 06 H200 NVL 141 GB · 4,890 GB/s · Q8_0 2,589 tok/s
  7. 07 H200 SXM 141 GB 141 GB · 4,890 GB/s · Q8_0 2,589 tok/s
  8. 08 Radeon Instinct MI325X 256 GB · 6,000 GB/s · Q8_0 2,478 tok/s
  9. 09 Radeon Instinct MI300A 128 GB · 5,325 GB/s · Q8_0 2,199 tok/s
  10. 10 Radeon Instinct MI300X 192 GB · 5,325 GB/s · Q8_0 2,199 tok/s

The smallest GPUs that still run MedSigLIP

The cheapest route in, by memory capacity. A tight fit runs the model but leaves nothing spare for a longer conversation.

  1. 01 GeForce RTX 4010 4 GB · needs 1.6 GB · Q8_0 · comfortable 50.8 tok/s
  2. 02 RTX A400 4 GB · needs 1.6 GB · Q8_0 · comfortable 50.8 tok/s
  3. 03 RTX 500 Mobile Ada Generation 4 GB · needs 1.6 GB · Q8_0 · comfortable 67.8 tok/s
  4. 04 GeForce RTX 3050 A Mobile 4 GB · needs 1.6 GB · Q8_0 · comfortable 102 tok/s
  5. 05 Jetson Orin Nano 4 GB 4 GB · needs 1.6 GB · Q8_0 · comfortable 18.1 tok/s
  6. 06 Radeon RX 6450M 4 GB · needs 1.6 GB · Q8_0 · comfortable 52.9 tok/s
  7. 07 Radeon RX 6550M 4 GB · needs 1.6 GB · Q8_0 · comfortable 59.5 tok/s
  8. 08 Radeon RX 6550S 4 GB · needs 1.6 GB · Q8_0 · comfortable 52.9 tok/s
  9. 09 Arc A310 4 GB · needs 1.6 GB · Q8_0 · comfortable 42.7 tok/s
  10. 10 Arc Pro A30M 4 GB · needs 1.6 GB · Q8_0 · comfortable 44.1 tok/s

What the numbers mean

The hardware side

Minimum card

Tesla C1080

Memory needed

1.6 GB

Fastest

4,235 tok/s

MedSigLIP is small enough at 800M parameters that hardware is rarely the obstacle — 818 of the cards we track can run it, including cards several years old.

At the low end, a Tesla C1080 handles it — 4 GB, at Q8_0, for about 46.1 tokens per second.

A B200 is the fastest we calculate for it: about 4,235 tokens per second, from 8,000 GB/s of memory bandwidth.

What this model is

MedSigLIP was published by Google, in United States of America, in July 2025. The organisation is categorised as industry.

It works in Vision, Medicine, and is recorded as doing image embedding, Image segmentation, Image classification.

Published weights mean the model runs on your machine rather than someone else's, which is what makes the hardware question below answerable at all. It is published under the google organisation on Hugging Face.

What decides the speed

Half the cards that hold it manage more than 118.9 tokens per second, and 809 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.

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.

Step by step

How to choose a GPU for MedSigLIP

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

  1. 01

    Start from the memory column

    The table lists every card that can hold MedSigLIP — around 1.6 GB at Q8_0. That figure, not the card's headline performance, is what decides whether it runs.

  2. 02

    Match the context to your actual use

    Longer conversations cost memory on top of what the weights need. Move the slider to your real working length before trusting any row for MedSigLIP.

  3. 03

    Choose how far you will compress it

    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 MedSigLIP by squeezing it further than you would want.

  4. 04

    Sort by speed

    Sort by speed to see how cards rank for MedSigLIP. It will not match a gaming ordering — generation is bound by memory bandwidth, which is why the B200 tops it at 4,235 tok/s.

  5. 05

    Look at the headroom, not just the fit

    Tight means MedSigLIP loads and works, with no room to raise the context later. Comfortable means you can. The difference matters more than a few tokens per second.

  6. 06

    Check the card from the other side

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

Answers

MedSigLIP — common questions

01

Would two GPUs run MedSigLIP faster?

A second card roughly doubles the memory available but not the generation rate. With 818 cards already able to run MedSigLIP alone, the case for pairing is weak.

02

Why does the quantisation differ between cards for MedSigLIP?

Because capacity varies, so does how hard MedSigLIP has to be squeezed — 1 distinct levels appear in the table above. Set a minimum quality to compare at one.

03

How accurate are these MedSigLIP speed estimates?

These are estimates with real error bars. The fastest result here, 2,541–6,776 tok/s on the B200, could reasonably land anywhere in its published range depending on which runtime you use.

04

What GPU do I need to run MedSigLIP?

The smallest card in our catalogue that holds MedSigLIP is the Tesla C1080, with 4 GB of memory. It runs the model at Q8_0 using about 1.6 GB, and produces roughly 46.1 tokens per second. 818 cards in total can run it.

05

How fast is MedSigLIP on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 4,235 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 MedSigLIP clear that.

06

How much VRAM does MedSigLIP 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.

07

Can I run MedSigLIP 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 789 tokens per second — a comfortable fit.

08

Can I run MedSigLIP 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 483 tokens per second — a comfortable fit.

09

Can I run MedSigLIP 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 598 tokens per second — a comfortable fit.

10

Can I run MedSigLIP 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 709 tokens per second — a comfortable fit.

11

Is MedSigLIP open source?

Its weights are published, so MedSigLIP 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.

12

How many parameters does MedSigLIP have?

MedSigLIP has 800M parameters. "MedSigLIP contains a 400M parameter vision encoder and 400M parameter text encoder,". 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.

13

Who created MedSigLIP?

MedSigLIP was published by Google, based in United States of America, categorised as industry.

14

When was MedSigLIP released?

MedSigLIP was published in July 2025.

15

What is MedSigLIP used for?

MedSigLIP works in Vision, Medicine, and is recorded as handling image embedding, Image segmentation, Image classification. 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.

16

Where can I download MedSigLIP?

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

17

Can I run MedSigLIP if it does not fit in my GPU?

It can be split between the card and system memory, but MedSigLIP generates painfully slowly that way. Nothing on this page assumes offloading.

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