InternVL2_5-78B TPS calculator

Open weights Shanghai AI Lab,SenseTime,Tsinghua University,Nanjing University,Fudan University,Chinese University of Hong Kong (CUHK),Shanghai Jiao Tong University 78.4B 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

58 of 818 cards that can run it

Smallest card that fits

Quadro RTX 8000

48 GB · Q3_K_M · 9.8 tok/s

Fastest card

B200

43.2 tok/s · 180 GB

Which GPUs can run InternVL2_5-78B?

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

58 cards match

Calculating

					Needs	Quantisation	Fit
43.2 tok/s 26–69 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	84.6 GB	Q8_0	Comfortable
43.2 tok/s 26–69 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	84.6 GB	Q8_0	Comfortable
34.5 tok/s 21–55 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	84.6 GB	Q8_0	Comfortable
34.5 tok/s 21–55 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	84.6 GB	Q8_0	Comfortable
30.9 tok/s 19–49 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	66.4 GB	Q6_K	Comfortable
27.6 tok/s 17–44 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	84.6 GB	Q8_0	Comfortable
27.3 tok/s 16–44 · low confidence	GRID A100B NVIDIA	48 GB	1,870 GB/s	May 2020	39.0 GB	Q3_K_M	Tight
26.4 tok/s 16–42 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	84.6 GB	Q8_0	Comfortable
26.4 tok/s 16–42 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	84.6 GB	Q8_0	Comfortable
26.4 tok/s 16–42 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	66.4 GB	Q6_K	Tight
26.4 tok/s 16–42 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	66.4 GB	Q6_K	Comfortable
26.4 tok/s 16–42 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	66.4 GB	Q6_K	Tight
25.3 tok/s 15–40 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	84.6 GB	Q8_0	Comfortable
22.4 tok/s 13–36 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	84.6 GB	Q8_0	Comfortable
22.4 tok/s 13–36 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	84.6 GB	Q8_0	Comfortable
22.4 tok/s 13–36 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	84.6 GB	Q8_0	Comfortable
19.5 tok/s 12–31 · low confidence	RTX PRO 5000 Blackwell NVIDIA	48 GB	1,340 GB/s	Mar 2025	39.0 GB	Q3_K_M	Tight
19.5 tok/s 12–31 · low confidence	H100 SXM5 64 GB NVIDIA	64 GB	2,020 GB/s	Mar 2023	57.3 GB	Q5_K_M	Tight
18.2 tok/s 11–29 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	84.6 GB	Q8_0	Tight
18.2 tok/s 11–29 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	84.6 GB	Q8_0	Tight
16.0 tok/s 10–26 · low confidence	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	66.4 GB	Q6_K	Tight
16.0 tok/s 10–26 · low confidence	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	66.4 GB	Q6_K	Tight
16.0 tok/s 10–26 · low confidence	A800 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Aug 2022	66.4 GB	Q6_K	Tight
16.0 tok/s 10–26 · low confidence	H100 CNX NVIDIA	80 GB	2,040 GB/s	Mar 2023	66.4 GB	Q6_K	Tight
16.0 tok/s 10–26 · low confidence	H100 PCIe 80 GB NVIDIA	80 GB	2,040 GB/s	Oct 2022	66.4 GB	Q6_K	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, Question answering, Character recognition (OCR)
Base model: Qwen2.5 Instruct (72B),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: 78.4B
Training data: 120,000,000,000 tokens
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: 2.1 × 10²² 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)
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.

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_5-78B

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 InternVL2_5-78B

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

Quadro RTX 8000

Memory needed

39.0 GB

Fastest

43.2 tok/s

InternVL2_5-78B sits at 78.4B parameters, which puts it above consumer hardware and into the range where a card is bought for this purpose rather than repurposed for it. 58 of the cards we track can hold it.

The entry point is the Quadro RTX 8000: 48 GB of memory, Q3_K_M compression, roughly 9.8 tokens per second.

Top of the range is the B200, at roughly 43.2 tokens per second thanks to 8,000 GB/s of bandwidth.

What this model is

InternVL2_5-78B 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. It comes out of academia,Industry,Academia,Academia,Academia,Academia,Academia.

It works in Multimodal, Language, Vision, and is recorded as doing visual question answering, Language modeling/generation, Question answering, Character recognition (OCR).

It builds on Qwen2.5 Instruct (72B),InternViT-6B, which is why it shares that model's general shape and size.

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 OpenGVLab organisation on Hugging Face.

What decides the speed

Across every card that can run it, the middle of the range is about 13.9 tokens per second, and 44 of them clear the ten tokens per second that roughly matches reading speed.

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.

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.

Training and provenance

It was trained on about 120,000,000,000 tokens of text.

Step by step

How to choose a GPU for InternVL2_5-78B

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

01

Start from the memory column

Look at what InternVL2_5-78B actually needs — around 39.0 GB at Q3_K_M. No amount of processing power compensates for a card that cannot hold it.
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 InternVL2_5-78B.
03

Set a quality floor

Compression is what makes InternVL2_5-78B fit smaller cards, at some cost in accuracy — Q3_K_M on the smallest card that fits. A minimum quality removes the ones that go too far.
04

Sort by speed

The speed ordering for InternVL2_5-78B is effectively an ordering by memory bandwidth, which is why the B200 tops it at 43.2 tok/s.
05

Read the fit column last

The fit column separates cards that just manage InternVL2_5-78B from those with room to spare. Buy for the second if the context might grow.
06

See what else that card runs

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_5-78B alone — a card is usually bought for more than one model.

Answers

InternVL2_5-78B — common questions

What is InternVL2_5-78B used for?

InternVL2_5-78B works in Multimodal, Language, Vision, and is recorded as handling visual question answering, Language modeling/generation, Question answering, Character recognition (OCR). 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.

Where can I download InternVL2_5-78B?

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.

Can I run InternVL2_5-78B 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 — the nearest miss we calculate is short by 12.1 GB. Our figures for InternVL2_5-78B assume it is fully resident.

Would two GPUs run InternVL2_5-78B faster?

Capacity adds across cards; throughput does not. Since 58 of the cards we track already hold InternVL2_5-78B on their own, a second card is rarely the answer here.

Why does the quantisation differ between cards for InternVL2_5-78B?

Because capacity varies, so does how hard InternVL2_5-78B has to be squeezed — 4 distinct levels appear in the table above. Set a minimum quality to compare at one.

How accurate are these InternVL2_5-78B speed estimates?

They are calculated from specifications rather than measured, and each carries a range — 26–69 tok/s on the B200, for instance. The same model and card vary by thirty to fifty per cent depending on the inference software and its version.

What GPU do I need to run InternVL2_5-78B?

The smallest card in our catalogue that holds InternVL2_5-78B is the Quadro RTX 8000, with 48 GB of memory. It runs the model at Q3_K_M using about 39.0 GB, and produces roughly 9.8 tokens per second. 58 cards in total can run it.

How fast is InternVL2_5-78B on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 43.2 tokens per second; the slowest that still runs it manages considerably less. Reading speed is around ten tokens per second, and 44 of the cards that can run InternVL2_5-78B clear that.

How much VRAM does InternVL2_5-78B need?

About 39.0 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.

Is InternVL2_5-78B open source?

Its weights are published, so InternVL2_5-78B 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 InternVL2_5-78B have?

InternVL2_5-78B has 78.4B parameters. 78.4B. 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.

Who created InternVL2_5-78B?

InternVL2_5-78B 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.

When was InternVL2_5-78B released?

InternVL2_5-78B 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.

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