InternVL2-40B TPS calculator

Open weights Shanghai AI Lab 40.1B parameters July 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 cards that can run it

818 cards we hold specifications for

Smallest card that fits

Tesla M40 24 GB

24 GB · Q3_K_M · 7.0 tok/s

Fastest card

B200

84.5 tok/s · 180 GB

Which GPUs can run InternVL2-40B?

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

126 cards match

Calculating

					Needs	Quantisation	Fit
84.5 tok/s 51–135 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	43.6 GB	Q8_0	Comfortable
84.5 tok/s 51–135 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	43.6 GB	Q8_0	Comfortable
67.5 tok/s 40–108 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	43.6 GB	Q8_0	Comfortable
67.5 tok/s 40–108 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	43.6 GB	Q8_0	Comfortable
54.0 tok/s 32–86 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	43.6 GB	Q8_0	Comfortable
51.7 tok/s 31–83 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	43.6 GB	Q8_0	Comfortable
51.7 tok/s 31–83 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	43.6 GB	Q8_0	Comfortable
49.4 tok/s 30–79 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	43.6 GB	Q8_0	Comfortable
45.6 tok/s 27–73 · low confidence	DRIVE A100 PROD NVIDIA	32 GB	1,870 GB/s	May 2020	25.0 GB	Q4_K_M	Tight
45.6 tok/s 27–73 · low confidence	GRID A100A NVIDIA	32 GB	1,870 GB/s	May 2020	25.0 GB	Q4_K_M	Tight
43.9 tok/s 26–70 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	43.6 GB	Q8_0	Comfortable
43.9 tok/s 26–70 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	43.6 GB	Q8_0	Comfortable
43.9 tok/s 26–70 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	43.6 GB	Q8_0	Comfortable
43.7 tok/s 26–70 · low confidence	GeForce RTX 5090 NVIDIA	32 GB	1,790 GB/s	Jan 2025	25.0 GB	Q4_K_M	Tight
43.7 tok/s 26–70 · low confidence	GeForce RTX 5090 D NVIDIA	32 GB	1,790 GB/s	Jan 2025	25.0 GB	Q4_K_M	Tight
41.6 tok/s 25–67 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	43.6 GB	Q8_0	Comfortable
38.2 tok/s 23–61 · low confidence	GeForce RTX 5090 D V2 NVIDIA	24 GB	1,340 GB/s	Aug 2025	20.3 GB	Q3_K_M	Tight
35.5 tok/s 21–57 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	43.6 GB	Q8_0	Comfortable
35.5 tok/s 21–57 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	43.6 GB	Q8_0	Comfortable
35.5 tok/s 21–57 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	43.6 GB	Q8_0	Comfortable
35.5 tok/s 21–57 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	43.6 GB	Q8_0	Comfortable
35.5 tok/s 21–57 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	43.6 GB	Q8_0	Comfortable
34.8 tok/s 21–56 · low confidence	A30X NVIDIA	24 GB	1,220 GB/s	Apr 2021	20.3 GB	Q3_K_M	Tight
28.8 tok/s 17–46 · low confidence	GeForce RTX 3090 Ti NVIDIA	24 GB	1,010 GB/s	Jan 2022	20.3 GB	Q3_K_M	Tight
28.8 tok/s 17–46 · low confidence	GeForce RTX 4090 NVIDIA	24 GB	1,010 GB/s	Sep 2022	20.3 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
Organisation type: Academia
Country: China
Published: 4 July 2024

What it does

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

Domain: Multimodal, Language, Vision, Video
Task: Visual question answering, Language modeling/generation, Question answering, Video description, Character recognition (OCR)
Base model: Nous-Hermes-2-Yi-34B,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: 40.1B
Training data: tokens
Epochs: 1

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.

Record confidence: Confident
Citations: 190

Sources

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

Reference: InternVL2: Better than the Best—Expanding Performance Boundaries of Open-Source Multimodal Models with the Progressive Scaling Strategy
Last updated: 28 November 2025

The extremes

The ten fastest GPUs that run InternVL2-40B

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 InternVL2-40B

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

What you need to run it

Minimum card

Tesla M40 24 GB

Memory needed

20.3 GB

Fastest

84.5 tok/s

InternVL2-40B reaches a parameter count of 40.1B. That lands in the range a serious desktop card can handle once the weights are compressed. The number of cards we track that can run it: 126.

The least hardware that works is Tesla M40 24 GB, with a memory capacity of 24 GB, running it at a compression of Q3_K_M and producing around 7.0 tokens per second.

The quickest result comes from B200, generating roughly 84.5 tokens per second on the strength of a memory bandwidth of 8,000 GB/s.

Background

InternVL2-40B was published by Shanghai AI Lab, in the country recorded as China, during July 2024. It comes out of an organisation categorised as academia.

It works in the domain of Multimodal, Language, Vision, Video, and is recorded as performing the task of visual question answering, Language modeling/generation, Question answering, Video description, Character recognition (OCR).

It builds on Nous-Hermes-2-Yi-34B,InternViT-6B. Most models at this scale are adapted from an existing base rather than built from nothing.

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. On Hugging Face it is published under the organisation OpenGVLab.

Reading the throughput figures

Half the cards that hold it manage more than 21.2 tokens per second. Producing text faster than most people read it: 106 of them.

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.

Step by step

How to choose a GPU for InternVL2-40B

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

Every card here has been checked against InternVL2-40B, needing around 20.3 GB at a compression of Q3_K_M. That figure, not the headline performance of a card, 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 a card that seemed fine stops fitting InternVL2-40B.
03

Decide how much compression you will accept

Each card runs the least-compressed copy it can hold, reaching a compression of Q3_K_M on the smallest card that fits. Setting a minimum quality drops the cards that only manage it by squeezing further than you would want, and holds the comparison at one level.
04

Compare tokens per second, not specifications

The speed ordering is effectively an ordering by memory bandwidth, for InternVL2-40B. It will not match a gaming ordering, because generation is bound by memory bandwidth. The card topping the list is B200, at 84.5 tok/s.
05

Read the fit column last

Tight means it loads and works with no room to raise the context later, in the case of InternVL2-40B. Comfortable means you can grow the context later. That difference matters more than a few tokens per second, so buy for comfortable if you expect to.
06

See what else that card runs

Each card page repeats this sweep for every model we hold, answering what else the hardware is good for beyond InternVL2-40B.

Answers

InternVL2-40B — common questions

InternVL2-40B— can I run it if it does not fit in my GPU?

Partly. Layers that do not fit sit in system memory and run at a fraction of the speed, so a mostly-offloaded model is rarely worth using. The nearest miss we calculate falls short by 7.0 GB. Every figure here assumes the whole model is resident on the card.

InternVL2-40B— would two GPUs run it faster?

A second card roughly doubles the memory available but not the generation rate. The number already able to run it alone: 126. So a second card is rarely the answer here.

InternVL2-40B— why does the quantisation differ between cards?

Each card is shown running the least-compressed copy it can hold. The number of distinct compression levels across the cards that fit it: 4. Bigger cards get the more accurate version, and the quality floor above pins the comparison to one level.

InternVL2-40B— how accurate are these speed estimates?

Every figure is derived from memory bandwidth and model size, not benchmarked, which is why each is published as a range rather than a single number. One example: 51–135 tok/s on B200. The same model and card vary by thirty to fifty per cent depending on the inference software and its version.

InternVL2-40B— what GPU do I need to run it?

The smallest card in our catalogue that holds it is Tesla M40 24 GB, with a memory capacity of 24 GB. It runs the model at a compression of Q3_K_M using about 20.3 GB, and produces roughly 7.0 tokens per second. The number of cards able to run it in total: 126.

InternVL2-40B— how fast is it on a GPU?

It depends on the card. The quickest we calculate is B200, at about 84.5 tokens per second; the slowest that still runs it manages considerably less. Reading speed is around ten tokens per second, and the number of cards clearing that: 106.

InternVL2-40B— how much VRAM does it need?

It needs about 20.3 GB at a compression of Q3_K_M, 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.

InternVL2-40B— can I run it on a GPU holding 24 GB?

Yes. The card GeForce RTX 5090 D V2, holding 24 GB, runs it at a compression of Q3_K_M, using about 20.3 GB and generating roughly 38.2 tokens per second. The fit is tight.

InternVL2-40B— is it open source?

Its weights are published, so it 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.

InternVL2-40B— how many parameters does it have?

It has a parameter count of 40.1B. 40.1B. 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.

InternVL2-40B— who created it?

It was published by Shanghai AI Lab, based in China, an organisation categorised as academia.

InternVL2-40B— when was it released?

It was published in July 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.

InternVL2-40B— what is it used for?

It works in the domain of Multimodal, Language, Vision, Video, and is recorded as handling the task of visual question answering, Language modeling/generation, Question answering, Video description, 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.

InternVL2-40B— where can I download it?

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

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