U-PaLM (540B)
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
- 20 October 2022
- Authors
- Yi Tay, Jason Wei, Hyung Won Chung, Vinh Q. Tran, David R. So, Siamak Shakeri, Xavier Garcia, Huaixiu Steven Zheng, Jinfeng Rao, Aakanksha Chowdhery, Denny Zhou, Donald Metzler, Slav Petrov, Neil Houlsby, Quoc V. Le, Mostafa Dehghani
What it does
The problem areas the model was built for. A model can carry several of each.
- Domain
- Language
- Task
- Language generation, Language modeling/generation, Question answering, Mathematical reasoning, Quantitative reasoning
- Base model
- PaLM (540B)
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
- 540B
- Training data
- 1,300,000,000 tokens
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
- 2.5 × 10²⁴ FLOP
- How it was established
- Comparison with other models
- Fine-tuning compute
- 4 × 10²¹ FLOP
"The total number of extra tokens we train on for the 540B model is approximately 1.3 Billion which constitutes 0.16% extra computation... Training an U-PaLM 540B model only consumes 512 TPUv4 chips and finishes in about 5 days which is considered to be lightweight." original PaLM was 2.527e+24. adding 0.16% is ~2.53e24
"The total number of extra tokens we train on for the 540B model is approximately 1.3 Billion which constitutes 0.16% extra computation... Training an U-PaLM 540B model only consumes 512 TPUv4 chips and finishes in about 5 days which is considered to be lightweight." PaLM was 2.5e24 0.16% of that is 4e21
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 v4
- Chips used
- 512
- Chip-hours
- 61,440
- Wall-clock time
- 120 hours
- Power draw
- 348.3 kW
5 days
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.
- Frontier model
- Yes
- Likely above 10²³ FLOP
- Yes
- Why it is tracked
- SOTA improvement
- Record confidence
- Confident
- Citations
- 76
Figure 3 "We show that U-PaLM 540B outperforms PaLM 540B on 21 out of 26 tasks. Given that PaLM is the SOTA language model on these tasks, this makes U-PaLM the new state-of-the-art on these tasks." performance improvement equivalent to 2x training efficiency: "Impressively, at 540B scale, we show an approximately 2x computational savings rate where U-PaLM achieves the same performance as the final PaLM 540B model at around half its computational budget "
Sources
Where this record came from and when it was last checked.
- Reference
- Transcending Scaling Laws with 0.1% Extra Compute
- Last updated
- 25 May 2026
What the numbers mean
What this model is
U-PaLM (540B) was published by Google, in United States of America, in October 2022. It comes out of industry.
It works in Language, and is recorded as doing language generation, Language modeling/generation, Question answering, Mathematical reasoning, Quantitative reasoning.
Its starting point was PaLM (540B) — most models at this scale are adapted from an existing base rather than built from nothing.
Because the weights are not available, none of the hardware figures elsewhere on this site apply to it.
Training and provenance
Producing it required around 2.5 × 10²⁴ FLOP of arithmetic, on Google TPU v4, which is a statement about the training budget rather than about inference.
It was trained on about 1,300,000,000 tokens of text.
Its inclusion criterion is sOTA improvement.
Answers
U-PaLM (540B) — common questions
How many parameters does U-PaLM (540B) have?
U-PaLM (540B) has 540B 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.
Who created U-PaLM (540B)?
U-PaLM (540B) was published by Google, based in United States of America, categorised as industry.
When was U-PaLM (540B) released?
U-PaLM (540B) was published in October 2022. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.
What is U-PaLM (540B) used for?
U-PaLM (540B) works in Language, and is recorded as handling language generation, Language modeling/generation, Question answering, Mathematical reasoning, Quantitative reasoning. Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.
How much compute was used to train U-PaLM (540B)?
Around 2.5 × 10²⁴ FLOP, on Google TPU v4. 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.
What GPU do I need to run U-PaLM (540B)?
None. U-PaLM (540B) 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.
Is U-PaLM (540B) open source?
No. U-PaLM (540B) has not had its weights published, so it exists only as a service controlled by its owner.
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.