AlphaGo Lee

Closed weights DeepMind January 2016

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
DeepMind
Organisation type
Industry
Country
United Kingdom of Great Britain and Northern Ireland
Published
27 January 2016
Authors
David Silver, Aja Huang, Chris J. Maddison, Arthur Guez, Laurent Sifre, George van den Driessche, Julian Schrittwieser, Ioannis Antonoglou, Veda Panneershelvam, Marc Lanctot, Sander Dieleman, Dominik Grewe, John Nham, Nal Kalchbrenner, Ilya Sutskever, Timothy Lillicrap, Madeleine Leach, Koray Kavukcuoglu, Thore Graepel, Demis Hassabis

What it does

The problem areas the model was built for. A model can carry several of each.

Domain
Games
Task
Go

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.

Training data
300,000,000 tokens

We trained the policy network pσ to classify positions according to expert moves played in the KGS data set. This data set contains 29.4 million positions from 160,000 games played by KGS 6 to 9 dan human players; 35.4% of the games are handicap games.

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 × 10²¹ FLOP

This number is pretty uncertain. I expect it to be right to around a factor of 3, at least compared to AlphaGo Fan. The architecture used was pretty much the same as AlphaGo Fan, but it was "trained for longer" and had around 5.33x the number of convolutional layers of AlphaGo Fan (256/48 = 5.33). The convolutional layers are the major contributor to the training compute, so I somewhat arbitrarily just multiply the compute for AlphaGo Fan by 5. Thus 3.8e20 * 5 = 1.9e21 Otherwise there has be…

How it was established
Comparison with other models

The training run

What it physically took to train: which chips, how many, for how long, and what that drew from the wall.

Wall-clock time
696 hours (29 days)

Training times are given for several components: - Policy network classifier: 3 weeks - Policy network RL: 1 day - Value network regression: 1 week - Rollout policy: "Similar to the policy network, the weights π of the rollout policy are trained from 8 million positions from human games on the Tygem server to maximize log likelihood by stochastic gradient descent. Rollouts execute at approximately 1,000 simulations per second per CPU thread on an empty board." could suggest (8M sims / 1000 sims/…

Compute cost
$22,207

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
Why it is tracked
Highly cited
Record confidence
Speculative
Citations
18,175

Sources

Where this record came from and when it was last checked.

Reference
Mastering the game of Go with deep neural networks and tree search
Last updated
1 January 2026

What the numbers mean

Where it came from

AlphaGo Lee was published by DeepMind, in United Kingdom of Great Britain and Northern Ireland, in January 2016. The organisation is categorised as industry.

It works in Games, and is recorded as doing go.

Its weights were never published, so it can only be reached through its provider. No graphics card changes that.

How it was trained

Training it took roughly 1.9 × 10²¹ FLOP of computation — a measure of what producing the model cost, not of how fast it answers.

It was trained on about 300,000,000 tokens of text.

The reason it appears in this catalogue at all is highly cited.

Answers

AlphaGo Lee — common questions

01

When was AlphaGo Lee released?

AlphaGo Lee was published in January 2016. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.

02

What is AlphaGo Lee used for?

AlphaGo Lee works in Games, and is recorded as handling go. These are the areas it was designed around; they describe intent rather than a hard boundary.

03

How much compute was used to train AlphaGo Lee?

Around 1.9 × 10²¹ FLOP. 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.

04

What GPU do I need to run AlphaGo Lee?

None. AlphaGo Lee 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.

05

Is AlphaGo Lee open source?

No. AlphaGo Lee has not had its weights published, so it exists only as a service controlled by its owner.

06

How many parameters does AlphaGo Lee have?

No parameter count has been published for AlphaGo Lee, which is why no memory or speed figure appears on this page.

07

Who created AlphaGo Lee?

AlphaGo Lee was published by DeepMind, based in United Kingdom of Great Britain and Northern Ireland, categorised as industry.

Source

Original publication

Record last updated 1 January 2026

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.