BiomedGPT (182M) TPS calculator

Open weights Lehigh University,University of Georgia,Samsung Research America,Harvard Medical School,University of Pennsylvania 182M parameters May 2023

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 · 203 tok/s

Fastest card

B200

18,617 tok/s · 180 GB

Which GPUs can run BiomedGPT (182M)?

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
18,617 tok/s 11,170–29,787 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	0.9 GB	Q8_0	Comfortable
18,617 tok/s 11,170–29,787 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	0.9 GB	Q8_0	Comfortable
14,866 tok/s 8,920–23,785 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	0.9 GB	Q8_0	Comfortable
14,866 tok/s 8,920–23,785 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	0.9 GB	Q8_0	Comfortable
11,889 tok/s 7,133–19,023 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	0.9 GB	Q8_0	Comfortable
11,379 tok/s 6,828–18,207 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	0.9 GB	Q8_0	Comfortable
11,379 tok/s 6,828–18,207 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	0.9 GB	Q8_0	Comfortable
10,891 tok/s 6,534–17,425 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	0.9 GB	Q8_0	Comfortable
9,666 tok/s 5,799–15,465 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	0.9 GB	Q8_0	Comfortable
9,666 tok/s 5,799–15,465 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	0.9 GB	Q8_0	Comfortable
9,666 tok/s 5,799–15,465 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	0.9 GB	Q8_0	Comfortable
9,169 tok/s 5,501–14,670 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	0.9 GB	Q8_0	Comfortable
7,819 tok/s 4,691–12,510 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	0.9 GB	Q8_0	Comfortable
7,819 tok/s 4,691–12,510 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	0.9 GB	Q8_0	Comfortable
7,819 tok/s 4,691–12,510 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	0.9 GB	Q8_0	Comfortable
7,819 tok/s 4,691–12,510 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	0.9 GB	Q8_0	Comfortable
7,819 tok/s 4,691–12,510 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	0.9 GB	Q8_0	Comfortable
5,954 tok/s 3,572–9,526 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	0.9 GB	Q8_0	Comfortable
5,954 tok/s 3,572–9,526 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	0.9 GB	Q8_0	Comfortable
4,961 tok/s 2,977–7,938 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	0.9 GB	Q8_0	Comfortable
4,855 tok/s 2,913–7,769 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	0.9 GB	Q8_0	Comfortable
4,747 tok/s 2,848–7,596 · low confidence	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	0.9 GB	Q8_0	Comfortable
4,747 tok/s 2,848–7,596 · low confidence	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	0.9 GB	Q8_0	Comfortable
4,747 tok/s 2,848–7,596 · low confidence	A800 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Aug 2022	0.9 GB	Q8_0	Comfortable
4,747 tok/s 2,848–7,596 · low confidence	H100 CNX NVIDIA	80 GB	2,040 GB/s	Mar 2023	0.9 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: Lehigh University,University of Georgia,Samsung Research America,Harvard Medical School,University of Pennsylvania
Organisation type: Academia,Academia,Industry,Academia,Academia
Country: United States of America
Published: 23 May 2023
Authors: Kai Zhang, Rong Zhou, Eashan Adhikarla, Zhiling Yan, Yixin Liu, Jun Yu, Zhengliang Liu, Xun Chen, Brian D. Davison, Hui Ren, Jing Huang, Chen Chen, Yuyin Zhou, Sunyang Fu, Wei Liu, Tianming Liu, Xiang Li, Yong Chen, Lifang He, James Zou, Quanzheng Li, Hongfang Liu, Lichao Sun

What it does

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

Domain: Language, Vision, Medicine
Task: Visual question answering, Medical diagnosis, Image captioning, Image classification, Text summarization, Language modeling/generation, Mortality prediction

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: 182M
Training data: 578,699,057 tokens

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 RTX A5000
Chips used: 10
Power draw: 4.6 kW

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 (unrestricted)
Training code: Open source

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: BiomedGPT: A Generalist Vision-Language Foundation Model for Diverse Biomedical Tasks
Last updated: 11 February 2026

The extremes

The ten fastest GPUs that run BiomedGPT (182M)

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 BiomedGPT (182M)

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

Tesla C1080

Memory needed

0.9 GB

Fastest

18,617 tok/s

BiomedGPT (182M) is small enough at 182M parameters that hardware is rarely the obstacle — 818 of the cards we track can run it, including cards several years old.

The entry point is the Tesla C1080: 4 GB of memory, Q8_0 compression, roughly 203 tokens per second.

A B200 is the fastest we calculate for it: about 18,617 tokens per second, from 8,000 GB/s of memory bandwidth.

Where it came from

BiomedGPT (182M) was published by Lehigh University,University of Georgia,Samsung Research America,Harvard Medical School,University of Pennsylvania, in United States of America, in May 2023. The organisation is categorised as academia,Academia,Industry,Academia,Academia.

It works in Language, Vision, Medicine, and is recorded as doing visual question answering, Medical diagnosis, Image captioning, Image classification, Text summarization, Language modeling/generation, Mortality prediction.

The weights are published, so it can be downloaded and run on your own hardware indefinitely, offline, with no account attached.

Understanding the speeds

Half the cards that hold it manage more than 522.8 tokens per second, and 818 exceed reading speed outright.

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

Without the attention layout on record, the memory column is an approximation. It is close enough to choose hardware by, and least reliable at long context.

What went into building it

Around 578,699,057 tokens went into training it.

Step by step

How to choose a GPU for BiomedGPT (182M)

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

The table lists every card that can hold BiomedGPT (182M) — around 0.9 GB at Q8_0. That figure, not the card's headline performance, is what decides whether it runs.
02

Set the context length you will work at

Longer conversations cost memory on top of what the weights need. Move the slider to your real working length before trusting any row for BiomedGPT (182M).
03

Choose how far you will compress it

The quantisation column varies by card, because a bigger card holds a more accurate copy of BiomedGPT (182M) — Q8_0 on the smallest card that fits. Set a floor to hold the comparison at one level.
04

Sort by speed

Ranking by tokens per second for BiomedGPT (182M) follows memory bandwidth, not core counts, which is why the B200 tops it at 18,617 tok/s.
05

Read the fit column last

Tight means BiomedGPT (182M) 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.
06

Open the card you have settled on

Every card name links to its own page, which runs the same calculation across the whole model catalogue. Worth a look before buying for BiomedGPT (182M) alone — a card is usually bought for more than one model.

Answers

BiomedGPT (182M) — common questions

Where can I download BiomedGPT (182M)?

The weights for BiomedGPT (182M) are published, though we do not hold a repository link for it. This site calculates hardware requirements rather than hosting model files.

Can I run BiomedGPT (182M) if it does not fit in my GPU?

It can be split between the card and system memory, but BiomedGPT (182M) generates painfully slowly that way. Nothing on this page assumes offloading.

Would two GPUs run BiomedGPT (182M) faster?

A second card roughly doubles the memory available but not the generation rate. With 818 cards already able to run BiomedGPT (182M) alone, the case for pairing is weak.

Why does the quantisation differ between cards for BiomedGPT (182M)?

A larger card holds a more accurate copy. Across the cards that run BiomedGPT (182M), 1 compression levels are used; the floor control above pins it to one.

How accurate are these BiomedGPT (182M) speed estimates?

These are estimates with real error bars. The fastest result here, 11,170–29,787 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 BiomedGPT (182M)?

The smallest card in our catalogue that holds BiomedGPT (182M) is the Tesla C1080, with 4 GB of memory. It runs the model at Q8_0 using about 0.9 GB, and produces roughly 203 tokens per second. 818 cards in total can run it.

How fast is BiomedGPT (182M) on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 18,617 tokens per second; the slowest that still runs it manages considerably less. Reading speed is around ten tokens per second, and 818 of the cards that can run BiomedGPT (182M) clear that.

How much VRAM does BiomedGPT (182M) need?

About 0.9 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 BiomedGPT (182M) on a 8 GB GPU?

Yes. A CMP 170HX 8 GB with 8 GB runs it at Q8_0, using about 0.9 GB and generating roughly 3,467 tokens per second — a comfortable fit.

Can I run BiomedGPT (182M) on a 12 GB GPU?

Yes. A GeForce RTX 3080 Ti with 12 GB runs it at Q8_0, using about 0.9 GB and generating roughly 2,123 tokens per second — a comfortable fit.

Can I run BiomedGPT (182M) on a 16 GB GPU?

Yes. A Tesla V100 SXM2 16 GB with 16 GB runs it at Q8_0, using about 0.9 GB and generating roughly 2,630 tokens per second — a comfortable fit.

Can I run BiomedGPT (182M) on a 24 GB GPU?

Yes. A GeForce RTX 5090 D V2 with 24 GB runs it at Q8_0, using about 0.9 GB and generating roughly 3,118 tokens per second — a comfortable fit.

Is BiomedGPT (182M) open source?

Its weights are published, so BiomedGPT (182M) 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 BiomedGPT (182M) have?

BiomedGPT (182M) has 182M 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 BiomedGPT (182M)?

BiomedGPT (182M) was published by Lehigh University,University of Georgia,Samsung Research America,Harvard Medical School,University of Pennsylvania, based in United States of America, categorised as academia,Academia,Industry,Academia,Academia.

When was BiomedGPT (182M) released?

BiomedGPT (182M) was published in May 2023. 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 BiomedGPT (182M) used for?

BiomedGPT (182M) works in Language, Vision, Medicine, and is recorded as handling visual question answering, Medical diagnosis, Image captioning, Image classification, Text summarization, Language modeling/generation, Mortality prediction. 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.

Source

Original publication

Record last updated 11 February 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