Sparse Wide GPT-3 Small
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
- Cerebras Systems
- Organisation type
- Industry
- Country
- United States of America
- Published
- 21 March 2023
- Authors
- Shreyas Saxena, Vithursan Thangarasa, Abhay Gupta, Sean Lie
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
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
- 1.3B
- Training data
- 2,500,000,000 tokens
- Epochs
- 1
GPT-3 Small: 125M At 90% sparsity → ~1.25B total parameters At 75% sparsity → 125M * ksw^2 = 125M *(1/(1-0.75)) = 500M total pramters (I assume only 125M are active)
2.5B (table 13)
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
- 1.9 × 10¹⁸ FLOP
- How it was established
- Operation counting
6 FLOP / parameter / token * 125* 10^6 active parameters * 2.5 * 10^9 tokens = 1.875e+18 FLOP __________________ in the Algorithmic progress paper the estimation was 8.84 × 10^19 FLOP, they assumed WT-103 dataset (not mentioned in the paper) and different number of parameters (1.3* 10^9 - I am unsure where it comes from)
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
- Cerebras CS-2
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
paper has repo but no code for sparse GPT-3: https://github.com/CerebrasResearch/Sparse-IFT
How it is classified
Labels the source dataset applies when tracking notable models, and how confident it is in the entry.
- Record confidence
- Speculative
- Citations
- 8
- Benchmark data
- Sparse Wide GPT-3 Small
Sources
Where this record came from and when it was last checked.
- Reference
- Sparse Iso-FLOP Transformations for Maximizing Training Efficiency
- Last updated
- 25 May 2026
What the numbers mean
Where it came from
Sparse Wide GPT-3 Small was published by Cerebras Systems, in the country recorded as United States of America, during March 2023. The category the publisher falls under is industry.
It works in the domain of Language, and is recorded as performing the task of language modeling/generation, Question answering.
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 1.9 × 10¹⁸ FLOP, on hardware recorded as Cerebras CS-2. That figure measures what producing the model cost, and has no bearing on how fast it answers.
The training set ran to roughly 2,500,000,000 tokens of text.
Answers
Sparse Wide GPT-3 Small — common questions
Sparse Wide GPT-3 Small— how much compute was used to train it?
Training consumed around 1.9 × 10¹⁸ FLOP, on hardware recorded as Cerebras CS-2. 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.
Sparse Wide GPT-3 Small— 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.
Sparse Wide GPT-3 Small— is it open source?
No. Its weights have not been published, so it exists only as a service controlled by its owner.
Sparse Wide GPT-3 Small— how many parameters does it have?
It has a parameter count of 1.3B. GPT-3 Small: 125M At 90% sparsity → ~1.25B total parameters At 75% sparsity → 125M * ksw^2 = 125M *(1/(1-0.75)) = 500M total pramters (I assume only 125M are active). 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.
Sparse Wide GPT-3 Small— who created it?
It was published by Cerebras Systems, based in United States of America, an organisation categorised as industry.
Sparse Wide GPT-3 Small— when was it released?
It was published in March 2023. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.
Sparse Wide GPT-3 Small— 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. These are the areas it was designed around; they describe intent rather than a hard boundary.
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.