CoAtNet

Closed weights Google,Google Research,Google Brain 2.4B parameters June 2021

No estimate

No hardware requirements for this model

The weights for this model have not been published, so it cannot be downloaded or run on your own hardware at any size. It is reachable only through its provider, and no graphics card changes that.

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
Google,Google Research,Google Brain
Organisation type
Industry,Industry,Industry
Country
United States of America
Published
9 June 2021
Authors
Zihang Dai, Hanxiao Liu, Quoc V. Le, Mingxing Tan

What it does

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

Domain
Vision
Task
Image classification

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
2.4B
Training data
88,779,000,000,000 tokens

Used JFT-3B (3 billion images), but not stated for how many epochs. Based on GPU time, training took 4.27e+22 FLOPs. Table 5 indicates 2.586e12 FLOPs per image. Since training is roughly 3x the FLOP cost of inference, implies inference on full training set took 1.42e22 FLOP Then # images trained over is around 1.42e22 / 2.586e12 = 5,491,105,955 So probably ~1.83 epochs on 3B images

Epochs
1.83

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
4.3 × 10²² FLOP

20.1K TPU-v3 core-days TPUs have two cores per chip, and a chip is 123 teraflop/s https://cloud.google.com/tpu/docs/system-architecture-tpu-vm#tpu_v3 123 teraflop/s * 20100/2 * 24 * 3600 * 0.4 (utilization assumption for non-language models) = 4.27e22

How it was established
Hardware

The training run

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

Training hardware
Google TPU v3
Chip-hours
10,050
Compute cost
$1,887

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
Closed — provider access only
Model access
Unreleased
Training code
Unreleased

How it is classified

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

Why it is tracked
SOTA improvement

"Notably, when we further scale up CoAtNet with JFT-3B, it achieves 90.88% top-1 accuracy on ImageNet, establishing a new state-of-the-art result."

Record confidence
Confident
Citations
1,589

Sources

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

Reference
CoAtNet: Marrying Convolution and Attention for All Data Sizes
Last updated
25 May 2026

What the numbers mean

What this model is

CoAtNet was published by Google,Google Research,Google Brain, in the country recorded as United States of America, during June 2021. The publishing organisation is categorised as industry,Industry,Industry.

It works in the domain of Vision, and is recorded as performing the task of image classification.

Its weights were never published, so it can only be reached through its provider. No graphics card changes that.

How it was trained

Training it took a computation budget of roughly 4.3 × 10²² FLOP, on hardware recorded as Google TPU v3. That figure measures what producing the model cost, and has no bearing on how fast it answers.

The training set ran to roughly 88,779,000,000,000 tokens of text.

Its inclusion criterion: sOTA improvement.

Answers

CoAtNet — common questions

01

CoAtNet— how many parameters does it have?

It has a parameter count of 2.4B. 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.

02

CoAtNet— who created it?

It was published by Google,Google Research,Google Brain, based in United States of America, an organisation categorised as industry,Industry,Industry.

03

CoAtNet— when was it released?

It was published in June 2021. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.

04

CoAtNet— what is it used for?

It works in the domain of Vision, and is recorded as handling the task of image classification. Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.

05

CoAtNet— how much compute was used to train it?

Training consumed around 4.3 × 10²² FLOP, on hardware recorded as Google TPU v3. 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.

06

CoAtNet— what GPU do I need to run it?

None. This is a closed model — its weights were never published, so it cannot be downloaded or run on your own hardware at any price. It is reachable only through its provider.

07

CoAtNet— is it open source?

No. Its weights have not been published, so it exists only as a service controlled by its owner.

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.