DataRater test model (1B)
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 DeepMind
- Organisation type
- Industry
- Country
- United States of America
- Published
- 23 May 2025
- Authors
- Dan A. Calian, Gregory Farquhar, Iurii Kemaev, Luisa M. Zintgraf, Matteo Hessel, Jeremy Shar, Junhyuk Oh, András György, Tom Schaul, Jeffrey Dean, Hado van Hasselt, David Silver
What it does
The problem areas the model was built for. A model can carry several of each.
- Domain
- Language
- Task
- Language modeling
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
- 1B
- Training data
- tokens
"A baseline 1B model is trained on the unfiltered dataset, while a second 1B model is trained analogously on the same dataset, but filtered by a DataRater."
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
- 1.2 × 10²⁰ FLOP
c = 6nd where n = 1B, d = 20B c = 6 * 1B * 20B = 1.2e20 FLOPs "All inner language models and DataRater models are based on the Chinchilla [Hoffmann et al., 2022] transformer architectures [Vaswani et al., 2017]." "Meta-training a DataRater requires approximately 58.4% of the FLOPS needed to train a single 1B LLM."
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 v6e Trillium
- Chips used
- 32
- Power draw
- 23.8 kW
How it is classified
Labels the source dataset applies when tracking notable models, and how confident it is in the entry.
- Record confidence
- Likely
- Citations
- 12
Sources
Where this record came from and when it was last checked.
- Reference
- DataRater: Meta-Learned Dataset Curation
- Last updated
- 25 May 2026
What the numbers mean
Where it came from
DataRater test model (1B) was published by Google DeepMind, in the country recorded as United States of America, during May 2025. The publishing organisation is categorised as industry.
It works in the domain of Language, and is recorded as performing the task of language modeling.
Because the weights are not available, none of the hardware figures elsewhere on this site apply to it.
What went into building it
Producing it required arithmetic totalling around 1.2 × 10²⁰ FLOP, on hardware recorded as Google TPU v6e Trillium. That figure measures what producing the model cost, and has no bearing on how fast it answers.
Answers
DataRater test model (1B) — common questions
DataRater test model (1B)— how many parameters does it have?
It has a parameter count of 1B. "A baseline 1B model is trained on the unfiltered dataset, while a second 1B model is trained analogously on the same dataset, but filtered by a DataRater.". 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.
DataRater test model (1B)— who created it?
It was published by Google DeepMind, based in United States of America, an organisation categorised as industry.
DataRater test model (1B)— when was it released?
It was published in May 2025.
DataRater test model (1B)— what is it used for?
It works in the domain of Language, and is recorded as handling the task of language modeling. These are the areas it was designed around; they describe intent rather than a hard boundary.
DataRater test model (1B)— how much compute was used to train it?
Training consumed around 1.2 × 10²⁰ FLOP, on hardware recorded as Google TPU v6e Trillium. 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.
DataRater test model (1B)— 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.
DataRater test model (1B)— is it open source?
The licensing was never recorded in our source data. We treat unstated licensing as closed, because an unrecorded licence is not one to rely on.
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.