InternVL2_5-38B TPS calculator

Open weights Shanghai AI Lab,SenseTime,Tsinghua University,Nanjing University,Fudan University,Chinese University of Hong Kong (CUHK),Shanghai Jiao Tong University 38.4B parameters December 2024

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

126 of 818 cards that can run it

Smallest card that fits

Tesla M40 24 GB

24 GB · Q3_K_M · 7.3 tok/s

Fastest card

B200

88.2 tok/s · 180 GB

Which GPUs can run InternVL2_5-38B?

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.

126 cards match

Calculating
Needs Quantisation Fit
88.2 tok/s

53–141 · low confidence

B200 NVIDIA 180 GB 8,000 GB/s Jan 2024 41.8 GB Q8_0 Comfortable
88.2 tok/s

53–141 · low confidence

B300 NVIDIA 288 GB 8,000 GB/s Sep 2025 41.8 GB Q8_0 Comfortable
70.5 tok/s

42–113 · low confidence

Radeon Instinct MI350X AMD 288 GB 8,190 GB/s Jan 2025 41.8 GB Q8_0 Comfortable
70.5 tok/s

42–113 · low confidence

Radeon Instinct MI355X AMD 288 GB 8,190 GB/s Jan 2025 41.8 GB Q8_0 Comfortable
56.4 tok/s

34–90 · low confidence

Radeon Instinct MI300 AMD 128 GB 6,550 GB/s Jan 2023 41.8 GB Q8_0 Comfortable
53.9 tok/s

32–86 · low confidence

H200 NVL NVIDIA 141 GB 4,890 GB/s Nov 2024 41.8 GB Q8_0 Comfortable
53.9 tok/s

32–86 · low confidence

H200 SXM 141 GB NVIDIA 141 GB 4,890 GB/s Nov 2024 41.8 GB Q8_0 Comfortable
51.6 tok/s

31–83 · low confidence

Radeon Instinct MI325X AMD 256 GB 6,000 GB/s Oct 2024 41.8 GB Q8_0 Comfortable
45.8 tok/s

27–73 · low confidence

Radeon Instinct MI300A AMD 128 GB 5,325 GB/s Dec 2023 41.8 GB Q8_0 Comfortable
45.8 tok/s

27–73 · low confidence

Radeon Instinct MI300X AMD 192 GB 5,325 GB/s Dec 2023 41.8 GB Q8_0 Comfortable
45.8 tok/s

27–73 · low confidence

Radeon Instinct MI308X AMD 192 GB 5,325 GB/s Dec 2023 41.8 GB Q8_0 Comfortable
43.5 tok/s

26–70 · low confidence

H100 NVL 94 GB NVIDIA 94 GB 3,940 GB/s Mar 2023 41.8 GB Q8_0 Comfortable
39.9 tok/s

24–64 · low confidence

GeForce RTX 5090 D V2 NVIDIA 24 GB 1,340 GB/s Aug 2025 19.5 GB Q3_K_M Tight
37.1 tok/s

22–59 · low confidence

H100 PCIe 96 GB NVIDIA 96 GB 3,360 GB/s Mar 2023 41.8 GB Q8_0 Comfortable
37.1 tok/s

22–59 · low confidence

H100 SXM5 80 GB NVIDIA 80 GB 3,360 GB/s Oct 2022 41.8 GB Q8_0 Comfortable
37.1 tok/s

22–59 · low confidence

H100 SXM5 94 GB NVIDIA 94 GB 3,360 GB/s Mar 2023 41.8 GB Q8_0 Comfortable
37.1 tok/s

22–59 · low confidence

H100 SXM5 96 GB NVIDIA 96 GB 3,360 GB/s Mar 2023 41.8 GB Q8_0 Comfortable
37.1 tok/s

22–59 · low confidence

H800 SXM5 NVIDIA 80 GB 3,360 GB/s Mar 2023 41.8 GB Q8_0 Comfortable
36.8 tok/s

22–59 · low confidence

DRIVE A100 PROD NVIDIA 32 GB 1,870 GB/s May 2020 28.4 GB Q5_K_M Tight
36.8 tok/s

22–59 · low confidence

GRID A100A NVIDIA 32 GB 1,870 GB/s May 2020 28.4 GB Q5_K_M Tight
36.3 tok/s

22–58 · low confidence

A30X NVIDIA 24 GB 1,220 GB/s Apr 2021 19.5 GB Q3_K_M Tight
35.3 tok/s

21–56 · low confidence

GeForce RTX 5090 NVIDIA 32 GB 1,790 GB/s Jan 2025 28.4 GB Q5_K_M Tight
35.3 tok/s

21–56 · low confidence

GeForce RTX 5090 D NVIDIA 32 GB 1,790 GB/s Jan 2025 28.4 GB Q5_K_M Tight
30.1 tok/s

18–48 · low confidence

GeForce RTX 3090 Ti NVIDIA 24 GB 1,010 GB/s Jan 2022 19.5 GB Q3_K_M Tight
30.1 tok/s

18–48 · low confidence

GeForce RTX 4090 NVIDIA 24 GB 1,010 GB/s Sep 2022 19.5 GB Q3_K_M Tight

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,SenseTime,Tsinghua University,Nanjing University,Fudan University,Chinese University of Hong Kong (CUHK),Shanghai Jiao Tong University
Organisation type
Academia,Industry,Academia,Academia,Academia,Academia,Academia
Country
China, Hong Kong
Published
6 December 2024
Authors
Zhe Chen, Weiyun Wang, Yue Cao, Yangzhou Liu, Zhangwei Gao, Erfei Cui, Jinguo Zhu, Shenglong Ye, Hao Tian, Zhaoyang Liu, Lixin Gu, Xuehui Wang, Qingyun Li, Yimin Ren, Zixuan Chen, Jiapeng Luo, Jiahao Wang, Tan Jiang, Bo Wang, Conghui He, Botian Shi, Xingcheng Zhang, Han Lv, Yi Wang, Wenqi Shao, Pei Chu, Zhongying Tu, Tong He, Zhiyong Wu, Huipeng Deng, Jiaye Ge, Kai Chen, Kaipeng Zhang, Limin Wang,…

What it does

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

Domain
Multimodal, Language, Vision
Task
Visual question answering, Language modeling/generation
Base model
Qwen2.5 Instruct (32B),InternViT-6B

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
38.4B

38.4B

Training data
151,000,000,000 tokens

Stage 1: 107B tokens Stage 2: 44B tokens 151B tokens total

Epochs
1

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.

How it was established
Operation counting
Fine-tuning compute
1 × 10²² FLOP

Stage 1: 107B tokens (MLP projector - 91.79M parameters https://internvl.github.io/blog/2024-12-05-InternVL-2.5) Stage 2: 44B tokens (full model) 6 FLOP / parameter / token * (91.79 * 10^6 parameters * 107 * 10^6 tokens + 38 * 10^9 parameters * 44 * 10^9 tokens) = 1.0032059e+22 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 (unrestricted)
Training code
Open source

mit license https://huggingface.co/OpenGVLab/InternVL2_5-38B "This project is released under the MIT License. This project uses the pre-trained Qwen2.5-32B-Instruct as a component, which is licensed under the Apache License 2.0." MIT license https://github.com/OpenGVLab/InternVL "Release training / evaluation code for InternVL2.5 series Support liger kernels to save GPU memory" is checked off

Hugging Face
OpenGVLab

How it is classified

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

Record confidence
Confident
Citations
1,468

Sources

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

Reference
Expanding Performance Boundaries of Open-Source Multimodal Models with Model, Data, and Test-Time Scaling
Last updated
25 May 2026

The extremes

The ten fastest GPUs for InternVL2_5-38B

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 88.2 tok/s
  2. 02 B200 180 GB · 8,000 GB/s · Q8_0 88.2 tok/s
  3. 03 Radeon Instinct MI350X 288 GB · 8,190 GB/s · Q8_0 70.5 tok/s
  4. 04 Radeon Instinct MI355X 288 GB · 8,190 GB/s · Q8_0 70.5 tok/s
  5. 05 Radeon Instinct MI300 128 GB · 6,550 GB/s · Q8_0 56.4 tok/s
  6. 06 H200 NVL 141 GB · 4,890 GB/s · Q8_0 53.9 tok/s
  7. 07 H200 SXM 141 GB 141 GB · 4,890 GB/s · Q8_0 53.9 tok/s
  8. 08 Radeon Instinct MI325X 256 GB · 6,000 GB/s · Q8_0 51.6 tok/s
  9. 09 Radeon Instinct MI300A 128 GB · 5,325 GB/s · Q8_0 45.8 tok/s
  10. 10 Radeon Instinct MI300X 192 GB · 5,325 GB/s · Q8_0 45.8 tok/s

The smallest GPUs that still run InternVL2_5-38B

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

  1. 01 Arc Pro B60 24 GB · needs 19.5 GB · Q3_K_M · tight 8.8 tok/s
  2. 02 GeForce RTX 5090 D V2 24 GB · needs 19.5 GB · Q3_K_M · tight 39.9 tok/s
  3. 03 RTX PRO 4000 Blackwell SFF 24 GB · needs 19.5 GB · Q3_K_M · tight 12.9 tok/s
  4. 04 GeForce RTX 5090 Mobile 24 GB · needs 19.5 GB · Q3_K_M · tight 26.7 tok/s
  5. 05 RTX PRO 4000 Blackwell 24 GB · needs 19.5 GB · Q3_K_M · tight 20.0 tok/s
  6. 06 GeForce RTX 4090 D 24 GB · needs 19.5 GB · Q3_K_M · tight 30.1 tok/s
  7. 07 RTX 4500 Ada Generation 24 GB · needs 19.5 GB · Q3_K_M · tight 12.9 tok/s
  8. 08 L4 24 GB · needs 19.5 GB · Q3_K_M · tight 8.9 tok/s
  9. 09 Radeon RX 7900 XTX 24 GB · needs 19.5 GB · Q3_K_M · tight 22.3 tok/s
  10. 10 L40 CNX 24 GB · needs 19.5 GB · Q3_K_M · tight 25.7 tok/s

What the numbers mean

What it takes to run this model

Minimum card

Tesla M40 24 GB

Memory needed

19.5 GB

Fastest

88.2 tok/s

With 38.4B parameters, InternVL2_5-38B lands in the range a serious desktop card can handle once the weights are compressed. 126 of the cards we track can run it.

The smallest card that holds it is the Tesla M40 24 GB with 24 GB, running it at Q3_K_M and producing around 7.3 tokens per second.

The quickest result comes from a B200 at around 88.2 tokens per second — its 8,000 GB/s of bandwidth is what buys that.

About this model

InternVL2_5-38B was published by Shanghai AI Lab,SenseTime,Tsinghua University,Nanjing University,Fudan University,Chinese University of Hong Kong (CUHK),Shanghai Jiao Tong University, in China, in December 2024. academia,Industry,Academia,Academia,Academia,Academia,Academia is the category the publisher falls under.

It works in Multimodal, Language, Vision, and is recorded as doing visual question answering, Language modeling/generation.

It is derived from Qwen2.5 Instruct (32B),InternViT-6B rather than trained from scratch, which is the usual way a specialised model is produced.

The weights are published, so it can be downloaded and run on your own hardware indefinitely, offline, with no account attached. It is published under the OpenGVLab organisation on Hugging Face.

How fast it runs, and why

Across every card that can run it, the middle of the range is about 19.7 tokens per second, and 92 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.

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.

How it was trained

It was trained on about 151,000,000,000 tokens of text.

Step by step

How to choose a GPU for InternVL2_5-38B

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

  1. 01

    Read the memory figure first

    Look at what InternVL2_5-38B actually needs — around 19.5 GB at Q3_K_M. No amount of processing power compensates for a card that cannot hold it.

  2. 02

    Decide how long your conversations run

    Set the context to what you will actually use. The cache grows with the conversation, and it is the usual reason InternVL2_5-38B stops fitting a card that seemed fine.

  3. 03

    Choose how far you will compress it

    Each card runs the least-compressed copy it can hold — Q3_K_M on the smallest card that fits. Setting a floor drops the cards that only manage InternVL2_5-38B by squeezing it further than you would want.

  4. 04

    Rank by throughput rather than spec sheet

    Ranking by tokens per second for InternVL2_5-38B follows memory bandwidth, not core counts, which is why the B200 tops it at 88.2 tok/s.

  5. 05

    Look at the headroom, not just the fit

    A tight fit runs InternVL2_5-38B but leaves nothing spare for a longer conversation; comfortable has headroom. If you expect to grow the context, buy for comfortable.

  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 InternVL2_5-38B alone — a card is usually bought for more than one model.

Answers

InternVL2_5-38B — common questions

01

How fast is InternVL2_5-38B on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 88.2 tokens per second; the slowest that still runs it manages considerably less. Reading speed is around ten tokens per second, and 92 of the cards that can run InternVL2_5-38B clear that.

02

How much VRAM does InternVL2_5-38B need?

About 19.5 GB at Q3_K_M 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.

03

Can I run InternVL2_5-38B on a 24 GB GPU?

Yes. A GeForce RTX 5090 D V2 with 24 GB runs it at Q3_K_M, using about 19.5 GB and generating roughly 39.9 tokens per second — a tight fit.

04

Is InternVL2_5-38B open source?

Its weights are published, so InternVL2_5-38B 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.

05

How many parameters does InternVL2_5-38B have?

InternVL2_5-38B has 38.4B parameters. 38.4B. 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.

06

Who created InternVL2_5-38B?

InternVL2_5-38B was published by Shanghai AI Lab,SenseTime,Tsinghua University,Nanjing University,Fudan University,Chinese University of Hong Kong (CUHK),Shanghai Jiao Tong University, based in China, categorised as academia,Industry,Academia,Academia,Academia,Academia,Academia.

07

When was InternVL2_5-38B released?

InternVL2_5-38B was published in December 2024. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.

08

What is InternVL2_5-38B used for?

InternVL2_5-38B works in Multimodal, Language, Vision, and is recorded as handling visual question answering, Language modeling/generation. These are the areas it was designed around; they describe intent rather than a hard boundary.

09

Where can I download InternVL2_5-38B?

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.

10

Can I run InternVL2_5-38B if it does not fit in my GPU?

It can be split between the card and system memory, but InternVL2_5-38B generates painfully slowly that way — the nearest miss we calculate is short by 5.9 GB. Nothing on this page assumes offloading.

11

Would two GPUs run InternVL2_5-38B faster?

Two cards buy memory rather than speed. That matters for InternVL2_5-38B only if one card cannot hold it — 126 can, so a second adds little.

12

Why does the quantisation differ between cards for InternVL2_5-38B?

Each card is shown running the least-compressed copy it can hold, and InternVL2_5-38B appears at 5 different compression levels across the cards that fit it. Bigger cards get the more accurate version.

13

How accurate are these InternVL2_5-38B speed estimates?

Every figure is derived from memory bandwidth and model size, not benchmarked. That is why each is published as a range such as 53–141 tok/s on the B200 rather than a single number.

14

What GPU do I need to run InternVL2_5-38B?

The smallest card in our catalogue that holds InternVL2_5-38B is the Tesla M40 24 GB, with 24 GB of memory. It runs the model at Q3_K_M using about 19.5 GB, and produces roughly 7.3 tokens per second. 126 cards in total can run it.

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