JEST-L++

Closed weights DeepMind June 2024

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
DeepMind
Organisation type
Industry
Country
United Kingdom of Great Britain and Northern Ireland
Published
25 June 2024
Authors
Talfan Evans, Nikhil Parthasarathy, Hamza Merzic, Olivier J. Henaff

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.

Training data
98,304,000,000,000 tokens

"Our default training configuration follows that of SigLIP [54], with a ViT-B/16 and Bert-B image-text dual encoder, training on WebLI for 3 billion examples with a batch size of 32k and the sigmoid-contrastive loss... We split training across 256 TPUv5e chips," according to page 14 of https://arxiv.org/pdf/2406.17711v1.

Batch size
32,000

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 × 10²¹ FLOP

See the right subfigure of Figure 1 on page 2 of https://arxiv.org/pdf/2406.17711v1.

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 v5e
Chips used
256
Power draw
113.7 kW

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

(a) Did you include the code, data, and instructions needed to reproduce the main experimental results (either in the supplemental material or as a URL)? [No] We do not explicitly provide the code for our experiments, however we detail in pseudocode the main components of our method (see algorithms 1 and A.1). Our model implementations and base experimental configuration were adopted from the open source big_vision codebase [3]. We do not provide the reference or downstream training datasets, bu…

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
30

Sources

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

Reference
Data curation via joint example selection further accelerates multimodal learning
Last updated
25 May 2026

What the numbers mean

Where it came from

JEST-L++ was published by DeepMind, in the country recorded as United Kingdom of Great Britain and Northern Ireland, during June 2024. It comes out of an organisation categorised as industry.

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

This is a closed model: the trained values stayed with whoever produced them, and there is no local version to run.

How it was trained

Training it took a computation budget of roughly 2 × 10²¹ FLOP, on hardware recorded as Google TPU v5e. That figure measures what producing the model cost, and has no bearing on how fast it answers.

The training set ran to roughly 98,304,000,000,000 tokens of text.

Answers

JEST-L++ — common questions

01

JEST-L++— how much compute was used to train it?

Training consumed around 2 × 10²¹ FLOP, on hardware recorded as Google TPU v5e. 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.

02

JEST-L++— 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.

03

JEST-L++— is it open source?

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

04

JEST-L++— how many parameters does it have?

No parameter count has been published for it, which is why no memory or speed figure appears on this page.

05

JEST-L++— who created it?

It was published by DeepMind, based in United Kingdom of Great Britain and Northern Ireland, an organisation categorised as industry.

06

JEST-L++— when was it released?

It was published in June 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.

07

JEST-L++— what is it used for?

It works in the domain of Vision, and is recorded as handling the task of image classification. These are the areas it was designed around; they describe intent rather than a hard boundary.

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.