Make-A-Scene

Closed weights Meta AI 4B parameters March 2022

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
Meta AI
Organisation type
Industry
Country
United States of America
Published
24 March 2022
Authors
Oran Gafni, Adam Polyak, Oron Ashual, Shelly Sheynin, Devi Parikh, Yaniv Taigman

What it does

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

Domain
Image generation
Task
Image generation, Text-to-image

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
4B

"Experiments were performed with a 4 billion parameter transformer, generating a sequence of 256 text tokens, 256 scene tokens, and 1024 image tokens, that are then decoded into an image with a resolution of 256 × 256 or 512 × 512 pixels (depending on the model of choice)"

Training data
267,386,880,000 tokens

"The scene-based transformer is trained on a union of CC12m [7], CC [51], and subsets of YFCC100m [55] and Redcaps [10], amounting to 35m text-image pairs." "The models were trained for a total of 170k iterations, with a batch size of 1024" "Experiments were performed with a 4 billion parameter transformer, generating a sequence of 256 text tokens, 256 scene tokens, and 1024 image tokens, that are then decoded into an image with a resolution of 256 × 256 or 512 × 512 pixels (depending on the m…

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
6.4 × 10²¹ FLOP

6 FLOP / parameter / token * 4 * 10^9 parameters * 267386880000 tokens [see dataset size notes] = 6.4172851e+21 FLOP

How it was established
Operation counting

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.

Why it is tracked
SOTA improvement

"FID is calculated over a subset of 30k images generated from the MS-COCO validation set text prompts with no reranking, and provided in Tab. 4.6. The evaluated models are divided into two groups: trained with and without (denoted as filtered) the MS-COCO training set. In both scenarios our model achieves the lowest FID."

Record confidence
Likely

Sources

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

Reference
Make-A-Scene: Scene-Based Text-to-Image Generation with Human Priors
Last updated
28 November 2025

What the numbers mean

Background

Make-A-Scene was published by Meta AI, in United States of America, in March 2022. industry is the category the publisher falls under.

It works in Image generation, and is recorded as doing image generation, Text-to-image.

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 6.4 × 10²¹ FLOP. That figure describes the cost of creating it and has no bearing on how quickly it generates text.

Around 267,386,880,000 tokens went into training it.

The reason it appears in this catalogue at all is sOTA improvement.

Answers

Make-A-Scene — common questions

01

When was Make-A-Scene released?

Make-A-Scene 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.

02

What is Make-A-Scene used for?

Make-A-Scene works in Image generation, and is recorded as handling image generation, Text-to-image. 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.

03

How much compute was used to train Make-A-Scene?

Around 6.4 × 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 Make-A-Scene?

None. Make-A-Scene 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 Make-A-Scene open source?

No. Make-A-Scene has not had its weights published, so it exists only as a service controlled by its owner.

06

How many parameters does Make-A-Scene have?

Make-A-Scene has 4B parameters. "Experiments were performed with a 4 billion parameter transformer, generating a sequence of 256 text tokens, 256 scene tokens, and 1024 image tokens, that are then decoded into an image with a resolution of 256 × 256 or 512 × 512 pixels (depending on the model of choice)". 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.

07

Who created Make-A-Scene?

Make-A-Scene was published by Meta AI, based in United States of America, categorised as industry.

Source

Original publication

Record last updated 28 November 2025

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.