Statement Curriculum Learning
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
- OpenAI
- Organisation type
- Industry
- Country
- United States of America
- Published
- 2 March 2022
- Authors
- Stanislas Polu, Jesse Michael Han, Kunhao Zheng, Mantas Baksys, Igor Babuschkin, Ilya Sutskever
What it does
The problem areas the model was built for. A model can carry several of each.
- Domain
- Language, Mathematics
- Task
- Automated theorem proving, Language modeling/generation, Question answering, Mathematical reasoning
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
- 774M
- Training data
- 372,000,000,000 tokens
Table on p12 gives WebMath dataset size in GB of code. Uncompressed code probably has a similar number of tokens per gigabyte as natural language text, on the order of 3e8 tokens per GB.
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
- 1.8 × 10²² FLOP
- How it was established
- Hardware
pretraining: 6 FLOP/parameter/token * 774000000 parameters * 372000000000 tokens = 1.727568e+21 FLOP unknown number of epochs -> "Likely" confidence expert iteration: 312000000000000 FLOP/GPU/sec * 48000 GPU-hours * 3600 sec / hour * 0.3 [assumed utilization] = 1.617408e+22 FLOP 1.617408e+22 FLOP + 1.727568e+21 FLOP = 1.7901648e+22 FLOP
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
- NVIDIA A100
- Chip-hours
- 48,000
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
- API access
- Training code
- Unreleased
It seems that only inference code is here, no model weights or pre-training code: https://github.com/openai/lean-gym Apache 2.0 "We present lean-gym’s API"
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
- Likely
- Citations
- 161
"by applying this expert iteration to a manually curated set of problem statements, we achieve state-of-the-art on the miniF2F benchmark, automatically solving multiple challenging problems drawn from high school olympiads."
Sources
Where this record came from and when it was last checked.
- Reference
- Formal Mathematics Statement Curriculum Learning
- Last updated
- 25 May 2026
What the numbers mean
Background
Statement Curriculum Learning was published by OpenAI, in the country recorded as United States of America, during March 2022. It comes out of an organisation categorised as industry.
It works in the domain of Language, Mathematics, and is recorded as performing the task of automated theorem proving, Language modeling/generation, Question answering, Mathematical reasoning.
Its weights were never published, so it can only be reached through its provider. No graphics card changes that.
Training and provenance
The training run consumed about 1.8 × 10²² FLOP, on hardware recorded as NVIDIA A100. That figure measures what producing the model cost, and has no bearing on how fast it answers.
The training set ran to roughly 372,000,000,000 tokens of text.
Its inclusion criterion: sOTA improvement.
Answers
Statement Curriculum Learning — common questions
Statement Curriculum Learning— who created it?
It was published by OpenAI, based in United States of America, an organisation categorised as industry.
Statement Curriculum Learning— when was it released?
It was published in March 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.
Statement Curriculum Learning— what is it used for?
It works in the domain of Language, Mathematics, and is recorded as handling the task of automated theorem proving, 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.
Statement Curriculum Learning— how much compute was used to train it?
Training consumed around 1.8 × 10²² FLOP, on hardware recorded as NVIDIA A100. 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.
Statement Curriculum Learning— 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.
Statement Curriculum Learning— is it open source?
No. Its weights have not been published, so it exists only as a service controlled by its owner.
Statement Curriculum Learning— how many parameters does it have?
It has a parameter count of 774M. 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.
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.