MetNet
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
- Organisation type
- Industry
- Country
- United States of America
- Published
- 24 March 2020
- Authors
- Casper Kaae Sønderby, Lasse Espeholt, Jonathan Heek, Mostafa Dehghani, Avital Oliver, Tim Salimans, Shreya Agrawal, Jason Hickey, Nal Kalchbrenner
What it does
The problem areas the model was built for. A model can carry several of each.
- Domain
- Earth science
- Task
- Weather forecasting
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
- 225M
- Training data
- 7,045,120,000 tokens
" In total this setting for MetNet has 225M parameters."
"we randomly extracted 13,717 test and validation samples and kept increasing the training set size until we observed no over-fitting at 1.72 million training samples"
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
- 9.5 × 10¹⁸ FLOP
- How it was established
- Operation counting
6 FLOP / parameter / token * 7045120000 updates [see training dataset size notes, "Likely" confidence] * 225000000 parameters = 9.510912e+18 FLOP check: they report training in 256 TPUs (TPU v3 most likely) 9.510912e+18 FLOP / (123000000000000 FLOP / sec / chip * 256 TPUs * 3600 sec / hour * 0.3 [assumed utilization]) = 0.27 hours (seems to be an underestimation) -> 'Speculative' confidence
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
- 256
- Power draw
- 235.4 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
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
- Record confidence
- Speculative
- Citations
- 345
"MetNet improves upon the current operational NWP system HRRR for up to 8 hours of lead time" ... "Numerical Weather Prediction is the most successful framework to perform medium- and longrange (up to 6 days with high confidence) forecast to date (Bauer et al., 2015)." "MetNet outperforms HRRR up to 400 to 480 minutes and outperforms a strong optical flow method and the persistence baseline throughout the 480 minute range."
Sources
Where this record came from and when it was last checked.
- Reference
- MetNet: A Neural Weather Model for Precipitation Forecasting
- Last updated
- 25 May 2026
What the numbers mean
What this model is
MetNet was published by Google, in the country recorded as United States of America, during March 2020. The publishing organisation is categorised as industry.
It works in the domain of Earth science, and is recorded as performing the task of weather forecasting.
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
The training run consumed about 9.5 × 10¹⁸ 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.
It was trained on a corpus of about 7,045,120,000 tokens of text.
The reason it appears in this catalogue at all: sOTA improvement.
Answers
MetNet — common questions
MetNet— what is it used for?
It works in the domain of Earth science, and is recorded as handling the task of weather forecasting. 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.
MetNet— how much compute was used to train it?
Training consumed around 9.5 × 10¹⁸ 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.
MetNet— 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.
MetNet— is it open source?
No. Its weights have not been published, so it exists only as a service controlled by its owner.
MetNet— how many parameters does it have?
It has a parameter count of 225M. " In total this setting for MetNet has 225M parameters.". 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.
MetNet— who created it?
It was published by Google, based in United States of America, an organisation categorised as industry.
MetNet— when was it released?
It was published in March 2020. 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.