Evo 2 40B TPS calculator

Open weights Arc Institute,Stanford University,NVIDIA,Liquid,University of California (UC) Berkeley,Goodfire,Columbia University,University of California San Francisco 40.3B parameters February 2025

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

84.1 tok/s · 180 GB

Which GPUs can run Evo 2 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
84.1 tok/s 50–135 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	43.8 GB	Q8_0	Comfortable
84.1 tok/s 50–135 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	43.8 GB	Q8_0	Comfortable
67.1 tok/s 40–107 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	43.8 GB	Q8_0	Comfortable
67.1 tok/s 40–107 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	43.8 GB	Q8_0	Comfortable
53.7 tok/s 32–86 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	43.8 GB	Q8_0	Comfortable
51.4 tok/s 31–82 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	43.8 GB	Q8_0	Comfortable
51.4 tok/s 31–82 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	43.8 GB	Q8_0	Comfortable
49.2 tok/s 30–79 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	43.8 GB	Q8_0	Comfortable
45.4 tok/s 27–73 · low confidence	DRIVE A100 PROD NVIDIA	32 GB	1,870 GB/s	May 2020	25.1 GB	Q4_K_M	Tight
45.4 tok/s 27–73 · low confidence	GRID A100A NVIDIA	32 GB	1,870 GB/s	May 2020	25.1 GB	Q4_K_M	Tight
43.7 tok/s 26–70 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	43.8 GB	Q8_0	Comfortable
43.7 tok/s 26–70 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	43.8 GB	Q8_0	Comfortable
43.7 tok/s 26–70 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	43.8 GB	Q8_0	Comfortable
43.4 tok/s 26–69 · low confidence	GeForce RTX 5090 NVIDIA	32 GB	1,790 GB/s	Jan 2025	25.1 GB	Q4_K_M	Tight
43.4 tok/s 26–69 · low confidence	GeForce RTX 5090 D NVIDIA	32 GB	1,790 GB/s	Jan 2025	25.1 GB	Q4_K_M	Tight
41.4 tok/s 25–66 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	43.8 GB	Q8_0	Comfortable
38.0 tok/s 23–61 · low confidence	GeForce RTX 5090 D V2 NVIDIA	24 GB	1,340 GB/s	Aug 2025	20.4 GB	Q3_K_M	Tight
35.3 tok/s 21–57 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	43.8 GB	Q8_0	Comfortable
35.3 tok/s 21–57 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	43.8 GB	Q8_0	Comfortable
35.3 tok/s 21–57 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	43.8 GB	Q8_0	Comfortable
35.3 tok/s 21–57 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	43.8 GB	Q8_0	Comfortable
35.3 tok/s 21–57 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	43.8 GB	Q8_0	Comfortable
34.6 tok/s 21–55 · low confidence	A30X NVIDIA	24 GB	1,220 GB/s	Apr 2021	20.4 GB	Q3_K_M	Tight
28.6 tok/s 17–46 · low confidence	GeForce RTX 3090 Ti NVIDIA	24 GB	1,010 GB/s	Jan 2022	20.4 GB	Q3_K_M	Tight
28.6 tok/s 17–46 · low confidence	GeForce RTX 4090 NVIDIA	24 GB	1,010 GB/s	Sep 2022	20.4 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: Arc Institute,Stanford University,NVIDIA,Liquid,University of California (UC) Berkeley,Goodfire,Columbia University,University of California San Francisco
Organisation type: Academia,Industry,Industry,Academia,Academia,Academia
Country: United States of America
Published: 19 February 2025
Authors: Garyk Brixi, Matthew G. Durrant, Jerome Ku, Michael Poli, Greg Brockman, Daniel Chang, Gabriel A. Gonzalez, Samuel H. King, David B. Li, Aditi T. Merchant, Mohsen Naghipourfar, Eric Nguyen, Chiara Ricci-Tam, David W. Romero, Gwanggyu Sun, Ali Taghibakshi, Anton Vorontsov, Brandon Yang, Myra Deng, Liv Gorton, Nam Nguyen, Nicholas K. Wang, Etowah Adams, Stephen A. Baccus, Steven Dillmann, Stefano Er…

What it does

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

Domain: Biology
Task: Protein or nucleotide language model (pLM/nLM)

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: 40.3B
Training data: 9,300,000,000,000 tokens

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: 2.3 × 10²⁴ FLOP
How it was established: Operation counting,Reported

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: arcinstitute

How it is classified

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

Likely above 10²³ FLOP: Yes
Record confidence: Confident

Sources

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

Reference: Genome modeling and design across all domains of life with Evo 2
Last updated: 28 November 2025

The extremes

The ten fastest GPUs that run Evo 2 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 Evo 2 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

The hardware side

Minimum card

Tesla M40 24 GB

Memory needed

20.4 GB

Fastest

84.1 tok/s

Evo 2 40B reaches a parameter count of 40.3B. 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.

At the low end it is handled by 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.

At the other end sits B200, generating roughly 84.1 tokens per second on the strength of a memory bandwidth of 8,000 GB/s.

What this model is

Evo 2 40B was published by Arc Institute,Stanford University,NVIDIA,Liquid,University of California (UC) Berkeley,Goodfire,Columbia University,University of California San Francisco, in the country recorded as United States of America, during February 2025. It comes out of an organisation categorised as academia,Industry,Industry,Academia,Academia,Academia.

It works in the domain of Biology, and is recorded as performing the task of protein or nucleotide language model (pLM/nLM).

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

What decides the speed

Half the cards that hold it manage more than 21.1 tokens per second. Producing text faster than most people read it: 106 of them.

It is a dense model, so every parameter is read for every token produced. That makes speed track memory bandwidth almost exactly — a card with twice the bandwidth generates roughly twice as fast.

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

The training run consumed about 2.3 × 10²⁴ FLOP. That figure measures what producing the model cost, and has no bearing on how fast it answers.

It was trained on a corpus of about 9,300,000,000,000 tokens of text.

Step by step

How to choose a GPU for Evo 2 40B

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

Every card here has been checked against Evo 2 40B, needing around 20.4 GB at a compression of Q3_K_M. No amount of processing power compensates for a card that cannot hold it.
02

Decide how long your conversations run

The conversation occupies memory too, and grows as it goes. Set the slider to the length you expect, because at long context a card that handles short questions easily can be dropped by Evo 2 40B.
03

Decide how much compression you will accept

The quantisation column varies by card, because a bigger card holds a more accurate copy, 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

Rank by throughput rather than spec sheet

Ranking by tokens per second follows memory bandwidth rather than core counts, for Evo 2 40B. It will not match a gaming ordering, because generation is bound by memory bandwidth. The card topping the list is B200, at 84.1 tok/s.
05

Read the fit column last

A tight fit runs, but leaves nothing spare for a longer conversation, in the case of Evo 2 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

Check the card from the other side

Following a card through to its own page shows every other model it can hold, which is the question that follows once you have settled on Evo 2 40B.

Answers

Evo 2 40B — common questions

Evo 2 40B— how much compute was used to train it?

Training consumed around 2.3 × 10²⁴ 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.

Evo 2 40B— can I run it if it does not fit in my GPU?

Only by offloading, which is usually a false economy: the part held in system memory drags the whole thing down. The nearest miss we calculate falls short by 7.1 GB. Every figure here assumes the whole model is resident on the card.

Evo 2 40B— would two GPUs run it faster?

Two cards buy memory rather than speed, which matters only if one card cannot hold it. The number that can: 126. So a second card is rarely the answer here.

Evo 2 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.

Evo 2 40B— how accurate are these speed estimates?

They are calculated from specifications rather than measured, and each carries a range. One example: 50–135 tok/s on B200. The same model and card vary by thirty to fifty per cent depending on the inference software and its version.

Evo 2 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 20.4 GB, and produces roughly 7.0 tokens per second. The number of cards able to run it in total: 126.

Evo 2 40B— how fast is it on a GPU?

It depends on the card. The quickest we calculate is B200, at about 84.1 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.

Evo 2 40B— how much VRAM does it need?

It needs about 20.4 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.

Evo 2 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 20.4 GB and generating roughly 38.0 tokens per second. The fit is tight.

Evo 2 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.

Evo 2 40B— how many parameters does it have?

It has a parameter count of 40.3B. Table 1 lists 40.3B parameters as model size. 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.

Evo 2 40B— who created it?

It was published by Arc Institute,Stanford University,NVIDIA,Liquid,University of California (UC) Berkeley,Goodfire,Columbia University,University of California San Francisco, based in United States of America, an organisation categorised as academia,Industry,Industry,Academia,Academia,Academia.

Evo 2 40B— when was it released?

It was published in February 2025.

Evo 2 40B— what is it used for?

It works in the domain of Biology, and is recorded as handling the task of protein or nucleotide language model (pLM/nLM). 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.

Evo 2 40B— where can I download it?

Its weights are published on Hugging Face, under the organisation arcinstitute. 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