Flan-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
- Hyung Won Chung, Le Hou, Shayne Longpre, Barret Zoph, Yi Tay, William Fedus, Yunxuan Li, Xuezhi Wang, Mostafa Dehghani, Siddhartha Brahma, Albert Webson, Shixiang Shane Gu, Zhuyun Dai, Mirac Suzgun, Xinyun Chen, Aakanksha Chowdhery, Alex Castro-Ros, Marie Pellat, Kevin Robinson, Dasha Valter, Sharan Narang, Gaurav Mishra, Adams Yu, Vincent Zhao, Yanping Huang, Andrew Dai, Hongkun Yu, Slav Petrov, …
What it does
The problem areas the model was built for. A model can carry several of each.
- Domain
- Language
- Task
- Language modeling/generation, Question answering, Mathematical 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
- 1,400,000,000 tokens
540B
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
- Reported,Hardware
- Fine-tuning compute
- 5.6 × 10²¹ FLOP
0.2% greater than Palm 540B, which used 2.5e24
5.6e21 per Table 2 "we only use 0.2% of the pre-training compute to instruction-finetune Flan-PaLM 540B (approximately 512 v4 TPU chips for 37 hours)" 512 * 37 * 3600 * 275 teraflops * 0.3 = 5.6e21 (so 30% utilization was correct)
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
- 18,944
- Wall-clock time
- 37 hours
- Hardware utilisation
- MFU 18.9%
- Power draw
- 348.3 kW
"we only use 0.2% of the pre-training compute to instruction-finetune Flan-PaLM 540B (approximately 512 v4 TPU chips for 37 hours)"
Estimated training compute: 2.5e24 FLOPs at 100% utilization, based on GPU-hours: 37 * 512 * 3600 * 3.12e14 = 2.128e22 Therefore, we can calculate utilization from known compute and maximum possible compute: MFU = 3.76e24 / 1.987e25 = 0.1892
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
- Highly cited,SOTA improvement
- Record confidence
- Confident
- Citations
- 4,129
>1k cites "Flan-PaLM 540B achieves state-of-the-art performance on several benchmarks, such as 75.2% on five-shot MMLU."
Sources
Where this record came from and when it was last checked.
- Reference
- Scaling Instruction-Finetuned Language Models
- Last updated
- 25 May 2026
What the numbers mean
Background
Flan-PaLM 540B was published by Google, in the country recorded as United States of America, during October 2022. It comes out of an organisation categorised as industry.
It works in the domain of Language, and is recorded as performing the task of language modeling/generation, Question answering, Mathematical reasoning.
Its starting point was an existing base model, PaLM (540B). That is the usual way a specialised model is produced.
Because the weights are not available, none of the hardware figures elsewhere on this site apply to it.
What went into building it
The training run consumed about 2.5 × 10²⁴ FLOP, on hardware recorded as Google TPU v4. That figure measures what producing the model cost, and has no bearing on how fast it answers.
Training consumed a corpus of around 1,400,000,000 tokens of text.
It is tracked in the underlying dataset for one reason in particular: highly cited,SOTA improvement.
Answers
Flan-PaLM 540B — common questions
Flan-PaLM 540B— when was it released?
It 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.
Flan-PaLM 540B— what is it used for?
It works in the domain of Language, and is recorded as handling the task of language modeling/generation, Question answering, Mathematical reasoning. 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.
Flan-PaLM 540B— how much compute was used to train it?
Training consumed around 2.5 × 10²⁴ FLOP, on hardware recorded as 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.
Flan-PaLM 540B— 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.
Flan-PaLM 540B— is it open source?
No. Its weights have not been published, so it exists only as a service controlled by its owner.
Flan-PaLM 540B— how many parameters does it have?
It has a parameter count of 540B. 540B. 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.
Flan-PaLM 540B— who created it?
It was published by Google, based in United States of America, an organisation categorised as industry.
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.