ALIGN
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 Research
- Organisation type
- Industry
- Country
- United States of America
- Published
- 11 June 2021
- Authors
- Chao Jia, Yinfei Yang, Ye Xia, Yi-Ting Chen, Zarana Parekh, Hieu Pham, Quoc V. Le, Yunhsuan Sung, Zhen Li, Tom Duerig
What it does
The problem areas the model was built for. A model can carry several of each.
- Domain
- Multimodal, Vision, Language
- Task
- Representation learning, Image classification, Image representation
- Approach
- Self-supervised learning
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
- 820M
- Training data
- 1,800,000,000 tokens
- Batch size
- 16,384
From author communication 480M (image tower) + 340 M (text tower)
Dataset contains 1.8B image-text pairs, then some duplicates are removed.
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
- 2.6 × 10²² FLOP
- How it was established
- Hardware
From author communication 14.82K TPUv3 core-days Precision: bfloat16 Estimation TPUv3 at float16: 123 TFLOPS/chip 123*10^12 TFLOPS/chip * (1 chip / 2 cores) * 14820 TPU core-days * 86400 s/day * 33% utilization = 2.599*10^22 FLOP https://www.wolframalpha.com/input?i=14820+days+*+123+teraFLOPS+%2F+2+*+0.33
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
- Chips used
- 512
- Chip-hours
- 177,818
- Wall-clock time
- 347 hours (14.5 days)
- Power draw
- 466.1 kW
- Compute cost
- $32,853
14820 TPU core-hours / 1024 TPU cores = 347.3 hours
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
- Highly cited,SOTA improvement
- Record confidence
- Confident
- Citations
- 5,463
"The aligned visual and language representations... set new state-of-the-art results on Flickr30K and MSCOCO image-text retrieval benchmarks"
Sources
Where this record came from and when it was last checked.
- Reference
- Scaling up visual and vision-language representation learning with noisy text supervision
- Last updated
- 25 May 2026
What the numbers mean
About this model
ALIGN was published by Google Research, in the country recorded as United States of America, during June 2021. The publishing organisation is categorised as industry.
It works in the domain of Multimodal, Vision, Language, and is recorded as performing the task of representation learning, Image classification, Image representation.
This is a closed model: the trained values stayed with whoever produced them, and there is no local version to run.
Training and provenance
Producing it required arithmetic totalling around 2.6 × 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.
Training consumed a corpus of around 1,800,000,000 tokens of text.
The reason it appears in this catalogue at all: highly cited,SOTA improvement.
Answers
ALIGN — common questions
ALIGN— what is it used for?
It works in the domain of Multimodal, Vision, Language, and is recorded as handling the task of representation learning, Image classification, Image representation. 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.
ALIGN— how much compute was used to train it?
Training consumed around 2.6 × 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.
ALIGN— 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.
ALIGN— is it open source?
No. Its weights have not been published, so it exists only as a service controlled by its owner.
ALIGN— how many parameters does it have?
It has a parameter count of 820M. From author communication 480M (image tower) + 340 M (text tower). 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.
ALIGN— who created it?
It was published by Google Research, based in United States of America, an organisation categorised as industry.
ALIGN— 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.
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.