Arctic TPS calculator

Open weights Snowflake 480B parameters April 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

3 cards that can run it

818 cards we hold specifications for

Smallest card that fits

B300

288 GB · Q3_K_M · 19.1 tok/s

Fastest card

B300

19.1 tok/s · 288 GB

Which GPUs can run Arctic?

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.

3 cards match

Calculating
Needs Quantisation Fit
19.1 tok/s

11–30 · low confidence

B300 NVIDIA 288 GB 8,000 GB/s Sep 2025 235.2 GB Q3_K_M Tight
15.2 tok/s

9–24 · low confidence

Radeon Instinct MI350X AMD 288 GB 8,190 GB/s Jan 2025 235.2 GB Q3_K_M Tight
15.2 tok/s

9–24 · low confidence

Radeon Instinct MI355X AMD 288 GB 8,190 GB/s Jan 2025 235.2 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
Snowflake
Organisation type
Industry
Country
United States of America
Published
24 April 2024
Authors
Snowflake AI Research

What it does

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

Domain
Language
Task
Language modeling/generation, Question answering, Code generation, Quantitative reasoning

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
480B

" It combines a 10B dense transformer model with a residual 128x3.66B MoE MLP resulting in 480B total and 17B active parameters chosen using a top-2 gating."

Training data
tokens

"Arctic was trained with a three-stage curriculum each with a different data composition focusing on generic skills in the first phase (1T Tokens), and enterprise-focused skills in the latter two phases (1.5T and 1T tokens). " 1+1.5+1 = 3.5

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
3.8 × 10²³ FLOP

from the graph: 1x - Arctic 1.9X - Llama 3 8B (7.2×10^23) ~ = Yi 34B (6.1e23) -> x = 3.2105263e+23 3X - Code Llama 70B (1.26e+24) -> x = 4.2e+23 17.5X - Llama 3 70B (7.861e24) -> x =4.492e+23 = 3.7975893e+23 Operation counting (17B active parameters): 6ND = 6 FLOP / parameter / token * 17*10^9 parameters * 3.5*10^12 tokens = 3.57e+23 FLOP geometric mean:(3.2105263e+23*4.492e+23*4.2e+23*3.57e+23)^(1/4) = 3.8347175e+23

How it was established
Other,Operation counting

The training run

What it physically took to train: which chips, how many, for how long, and what that drew from the wall.

Chip-hours
504,000
Compute cost
$2,000,000
Cloud vendor
AWS

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

Apache 2.0 license with ungated access to weights and code paired with open data recipe and research insights.

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
Snowflake Arctic: The Best LLM for Enterprise AI — Efficiently Intelligent, Truly Open
Last updated
28 November 2025

The extremes

The ten fastest GPUs that run Arctic

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 · Q3_K_M 19.1 tok/s
  2. 02 Radeon Instinct MI350X 288 GB · 8,190 GB/s · Q3_K_M 15.2 tok/s
  3. 03 Radeon Instinct MI355X 288 GB · 8,190 GB/s · Q3_K_M 15.2 tok/s

The smallest GPUs that still run Arctic

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

  1. 01 B300 288 GB · needs 235.2 GB · Q3_K_M · tight 19.1 tok/s
  2. 02 Radeon Instinct MI350X 288 GB · needs 235.2 GB · Q3_K_M · tight 15.2 tok/s
  3. 03 Radeon Instinct MI355X 288 GB · needs 235.2 GB · Q3_K_M · tight 15.2 tok/s

What the numbers mean

What it takes to run this model

Minimum card

B300

Memory needed

235.2 GB

Fastest

19.1 tok/s

Arctic reaches a parameter count of 480B. That is beyond what any single graphics card holds. Running it means either splitting it across several cards or renting hardware built for the job, and every card able to hold it alone is a datacentre part. The number that can: 3.

The smallest card that holds it is B300, with a memory capacity of 288 GB, running it at a compression of Q3_K_M and producing around 19.1 tokens per second.

Top of the range is B300, generating roughly 19.1 tokens per second on the strength of a memory bandwidth of 8,000 GB/s.

What this model is

Arctic was published by Snowflake, in the country recorded as United States of America, during April 2024. It comes out of an organisation categorised as industry.

It works in the domain of Language, and is recorded as performing the task of language modeling/generation, Question answering, Code generation, Quantitative reasoning.

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.

What decides the speed

Half the cards that hold it manage more than 15.2 tokens per second. Exceeding reading speed outright: 3 of them.

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

Producing it required arithmetic totalling around 3.8 × 10²³ FLOP. That figure measures what producing the model cost, and has no bearing on how fast it answers.

Step by step

How to choose a GPU for Arctic

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

  1. 01

    Start from the memory column

    The table lists every card able to hold Arctic, needing around 235.2 GB at a compression of Q3_K_M. That figure, not the headline performance of a card, is what decides whether it runs.

  2. 02

    Set the context length you will work at

    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 Arctic.

  3. 03

    Choose how far you will compress it

    Compression is what makes a model fit smaller cards, at some cost in accuracy, 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.

  4. 04

    Compare tokens per second, not specifications

    The speed ordering is effectively an ordering by memory bandwidth, for Arctic. It will not match a gaming ordering, because generation is bound by memory bandwidth. The card topping the list is B300, at 19.1 tok/s.

  5. 05

    Look at the headroom, not just the fit

    The fit column separates cards that just manage it from those with room to spare, in the case of Arctic. 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.

  6. 06

    Check the card from the other side

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

Answers

Arctic — common questions

01

Arctic— what is it used for?

It works in the domain of Language, and is recorded as handling the task of language modeling/generation, Question answering, Code generation, Quantitative reasoning. Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.

02

Arctic— where can I download it?

The weights are published, though we do not hold a repository link for it. This site calculates hardware requirements rather than hosting model files.

03

Arctic— how much compute was used to train it?

Training consumed around 3.8 × 10²³ 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.

04

Arctic— can I run it if it does not fit in my GPU?

It can be split between the card and system memory, but it generates painfully slowly that way. The nearest miss we calculate falls short by 60.7 GB. Every figure here assumes the whole model is resident on the card.

05

Arctic— 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: 3. So a second card is rarely the answer here.

06

Arctic— 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: 1. Bigger cards get the more accurate version, and the quality floor above pins the comparison to one level.

07

Arctic— how accurate are these speed estimates?

They are calculated from specifications rather than measured, and each carries a range. One example: 11–30 tok/s on B300. The same model and card vary by thirty to fifty per cent depending on the inference software and its version.

08

Arctic— what GPU do I need to run it?

The smallest card in our catalogue that holds it is B300, with a memory capacity of 288 GB. It runs the model at a compression of Q3_K_M using about 235.2 GB, and produces roughly 19.1 tokens per second. The number of cards able to run it in total: 3.

09

Arctic— how fast is it on a GPU?

It depends on the card. The quickest we calculate is B300, at about 19.1 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: 3.

10

Arctic— how much VRAM does it need?

It needs about 235.2 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.

11

Arctic— 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.

12

Arctic— how many parameters does it have?

It has a parameter count of 480B. " It combines a 10B dense transformer model with a residual 128x3.66B MoE MLP resulting in 480B total and 17B active parameters chosen using a top-2 gating.". 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.

13

Arctic— who created it?

It was published by Snowflake, based in United States of America, an organisation categorised as industry.

14

Arctic— when was it released?

It was published in April 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 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.