InternImage TPS calculator

Open weights Shanghai AI Lab,Tsinghua University,Nanjing University,SenseTime,Chinese University of Hong Kong (CUHK) 1.1B parameters November 2022

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 · 34.1 tok/s

Fastest card

B200

3,137 tok/s · 180 GB

Which GPUs can run InternImage?

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
3,137 tok/s 1,882–5,020 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	1.9 GB	Q8_0	Comfortable
3,137 tok/s 1,882–5,020 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	1.9 GB	Q8_0	Comfortable
2,505 tok/s 1,503–4,008 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	1.9 GB	Q8_0	Comfortable
2,505 tok/s 1,503–4,008 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	1.9 GB	Q8_0	Comfortable
2,004 tok/s 1,202–3,206 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	1.9 GB	Q8_0	Comfortable
1,918 tok/s 1,151–3,068 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	1.9 GB	Q8_0	Comfortable
1,918 tok/s 1,151–3,068 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	1.9 GB	Q8_0	Comfortable
1,835 tok/s 1,101–2,936 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	1.9 GB	Q8_0	Comfortable
1,629 tok/s 977–2,606 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	1.9 GB	Q8_0	Comfortable
1,629 tok/s 977–2,606 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	1.9 GB	Q8_0	Comfortable
1,629 tok/s 977–2,606 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	1.9 GB	Q8_0	Comfortable
1,545 tok/s 927–2,472 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	1.9 GB	Q8_0	Comfortable
1,318 tok/s 791–2,108 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	1.9 GB	Q8_0	Comfortable
1,318 tok/s 791–2,108 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	1.9 GB	Q8_0	Comfortable
1,318 tok/s 791–2,108 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	1.9 GB	Q8_0	Comfortable
1,318 tok/s 791–2,108 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	1.9 GB	Q8_0	Comfortable
1,318 tok/s 791–2,108 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	1.9 GB	Q8_0	Comfortable
1,003 tok/s 602–1,605 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	1.9 GB	Q8_0	Comfortable
1,003 tok/s 602–1,605 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	1.9 GB	Q8_0	Comfortable
836 tok/s 502–1,338 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	1.9 GB	Q8_0	Comfortable
818 tok/s 491–1,309 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	1.9 GB	Q8_0	Comfortable
800 tok/s 480–1,280 · low confidence	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	1.9 GB	Q8_0	Comfortable
800 tok/s 480–1,280 · low confidence	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	1.9 GB	Q8_0	Comfortable
800 tok/s 480–1,280 · low confidence	A800 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Aug 2022	1.9 GB	Q8_0	Comfortable
800 tok/s 480–1,280 · low confidence	H100 CNX NVIDIA	80 GB	2,040 GB/s	Mar 2023	1.9 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: Shanghai AI Lab,Tsinghua University,Nanjing University,SenseTime,Chinese University of Hong Kong (CUHK)
Organisation type: Academia,Academia,Academia,Industry,Academia
Country: China, Hong Kong
Published: 10 November 2022
Authors: Wenhai Wang, Jifeng Dai, Zhe Chen, Zhenhang Huang, Zhiqi Li, Xizhou Zhu, Xiaowei Hu, Tong Lu, Lewei Lu, Hongsheng Li, Xiaogang Wang, Yu Qiao

What it does

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

Domain: Vision
Task: Image classification, Object detection, Image segmentation
Numerical format: FP16

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: 1.1B
Training data: 83,692,000,000 tokens
Epochs: 30

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: 2.4 × 10²¹ 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 (unrestricted)
Training code: Open source
Hugging Face: OpenGVLab

How it is classified

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

Why it is tracked: SOTA improvement
Record confidence: Confident
Citations: 1,087

Sources

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

Reference: InternImage: Exploring Large-Scale Vision Foundation Models with Deformable Convolutions
Last updated: 25 May 2026

The extremes

The ten fastest GPUs that run InternImage

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 InternImage

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

The hardware side

Minimum card

Tesla C1080

Memory needed

1.9 GB

Fastest

3,137 tok/s

InternImage is small enough at 1.1B 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 34.1 tokens per second.

Top of the range is the B200, at roughly 3,137 tokens per second thanks to 8,000 GB/s of bandwidth.

Where it came from

InternImage was published by Shanghai AI Lab,Tsinghua University,Nanjing University,SenseTime,Chinese University of Hong Kong (CUHK), in China, in November 2022. It comes out of academia,Academia,Academia,Industry,Academia.

It works in Vision, and is recorded as doing image classification, Object detection, 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. It is published under the OpenGVLab organisation on Hugging Face.

Understanding the speeds

The median result is around 88.1 tokens per second; 799 cards produce text faster than most people read it.

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

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.

Training and provenance

Training it took roughly 2.4 × 10²¹ FLOP of computation — a measure of what producing the model cost, not of how fast it answers.

It was trained on about 83,692,000,000 tokens of text.

The reason it appears in this catalogue at all is sOTA improvement.

Step by step

How to choose a GPU for InternImage

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

The table lists every card that can hold InternImage — around 1.9 GB at Q8_0. That figure, not the card's headline performance, is what decides whether it runs.
02

Match the context to your actual use

Set the context to what you will actually use. The cache grows with the conversation, and it is the usual reason InternImage stops fitting a card that seemed fine.
03

Set a quality floor

The quantisation column varies by card, because a bigger card holds a more accurate copy of InternImage — 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 InternImage follows memory bandwidth, not core counts, which is why the B200 tops it at 3,137 tok/s.
05

Read the fit column last

Tight means InternImage 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

See what else that card runs

Each card page repeats this sweep for every model we hold. It answers what else the hardware is good for, beyond InternImage.

Answers

InternImage — common questions

Is InternImage open source?

Its weights are published, so InternImage 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 InternImage have?

InternImage has 1.1B parameters. 1.08B, table 1. 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 InternImage?

InternImage was published by Shanghai AI Lab,Tsinghua University,Nanjing University,SenseTime,Chinese University of Hong Kong (CUHK), based in China, categorised as academia,Academia,Academia,Industry,Academia.

When was InternImage released?

InternImage was published in November 2022. 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 InternImage used for?

InternImage works in Vision, and is recorded as handling image classification, Object detection, Image segmentation. Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.

Where can I download InternImage?

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

How much compute was used to train InternImage?

Around 2.4 × 10²¹ 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 InternImage if it does not fit in my GPU?

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

Would two GPUs run InternImage faster?

Capacity adds across cards; throughput does not. Since 818 of the cards we track already hold InternImage on their own, a second card is rarely the answer here.

Why does the quantisation differ between cards for InternImage?

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

How accurate are these InternImage speed estimates?

These are estimates with real error bars. The fastest result here, 1,882–5,020 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 InternImage?

The smallest card in our catalogue that holds InternImage is the Tesla C1080, with 4 GB of memory. It runs the model at Q8_0 using about 1.9 GB, and produces roughly 34.1 tokens per second. 818 cards in total can run it.

How fast is InternImage on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 3,137 tokens per second; the slowest that still runs it manages considerably less. Reading speed is around ten tokens per second, and 799 of the cards that can run InternImage clear that.

How much VRAM does InternImage need?

About 1.9 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 InternImage on a 8 GB GPU?

Yes. A CMP 170HX 8 GB with 8 GB runs it at Q8_0, using about 1.9 GB and generating roughly 584 tokens per second — a comfortable fit.

Can I run InternImage on a 12 GB GPU?

Yes. A GeForce RTX 3080 Ti with 12 GB runs it at Q8_0, using about 1.9 GB and generating roughly 358 tokens per second — a comfortable fit.

Can I run InternImage on a 16 GB GPU?

Yes. A Tesla V100 SXM2 16 GB with 16 GB runs it at Q8_0, using about 1.9 GB and generating roughly 443 tokens per second — a comfortable fit.

Can I run InternImage on a 24 GB GPU?

Yes. A GeForce RTX 5090 D V2 with 24 GB runs it at Q8_0, using about 1.9 GB and generating roughly 525 tokens per second — a comfortable fit.

Source

Original publication

Record last updated 25 May 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