DALL-E mini
No estimate
No hardware requirements for this model
This model's weights are open, but no parameter count has been published for it. Every memory and speed figure starts from that number, so we would rather show nothing than a fabricated estimate.
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
- Craiyon
- Organisation type
- Industry
- Country
- United States of America
- Published
- 26 October 2021
- Authors
- Dayma, Boris and Patil, Suraj and Cuenca, Pedro and Saifullah, Khalid and Abraham, Tanishq and Lê Khắc, Phúc and Melas, Luke and Ghosh, Ritobrata
What it does
The problem areas the model was built for. A model can carry several of each.
- Domain
- Vision
- Task
- 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.
- Training data
- 4,352,000,000 tokens
3M+12M+15M = 30M from https://huggingface.co/dalle-mini/dalle-mini#training-data
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
- 3.8 × 10¹⁹ FLOP
- How it was established
- Hardware
trained on a TPU v3-8 for 72 hours. That's 8 TPUv3 cores, or 4 TPUv3 chips (123 teraflop/s each) flops = (4) * (123 * 10**12) * (72 * 3600) * (0.3) = 3.8e19 (num gpu) * (peak flops) * (time in seconds) * (assumed utilization rate) from https://huggingface.co/dalle-mini/dalle-mini#dall%C2%B7e-mini-estimated-emissions
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 v3
- Chips used
- 4
- Chip-hours
- 288
- Wall-clock time
- 72 hours
- Power draw
- 3.6 kW
from https://huggingface.co/dalle-mini/dalle-mini#dall%C2%B7e-mini-estimated-emissions
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
- Open — downloadable
- Model access
- Open weights (unrestricted)
- Training code
- Open source
Apache 2.0 train code here: https://github.com/borisdayma/dalle-mini/blob/main/tools/train/train.py data is open with various licenses
How it is classified
Labels the source dataset applies when tracking notable models, and how confident it is in the entry.
- Record confidence
- Likely
Sources
Where this record came from and when it was last checked.
- Reference
- DALL·E Mini Model Card
- Last updated
- 28 November 2025
What the numbers mean
About this model
DALL-E mini was published by Craiyon, in United States of America, in October 2021. The organisation is categorised as industry.
It works in Vision, and is recorded as doing text-to-image.
The weights are published, so it can be downloaded and run on your own hardware indefinitely, offline, with no account attached.
What went into building it
Training it took roughly 3.8 × 10¹⁹ FLOP of computation, on Google TPU v3 — a measure of what producing the model cost, not of how fast it answers.
Around 4,352,000,000 tokens went into training it.
Answers
DALL-E mini — common questions
When was DALL-E mini released?
DALL-E mini was published in October 2021. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.
What is DALL-E mini used for?
DALL-E mini works in Vision, and is recorded as handling 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.
Where can I download DALL-E mini?
The weights for DALL-E mini are published, though we do not hold a repository link for it. This site calculates hardware requirements rather than hosting model files.
How much compute was used to train DALL-E mini?
Around 3.8 × 10¹⁹ FLOP, on Google TPU v3. 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.
What GPU do I need to run DALL-E mini?
We cannot say. DALL-E mini has open weights, but no parameter count has been published for it, and every memory and speed calculation starts from that number. We would rather show nothing than a fabricated estimate.
Is DALL-E mini open source?
Its weights are published, so DALL-E mini can be downloaded and run on your own hardware. Note that open weights is not the same as open source in the full sense — it says nothing about the training data, the training code, or the commercial terms attached.
How many parameters does DALL-E mini have?
No parameter count has been published for DALL-E mini, which is why no memory or speed figure appears on this page.
Who created DALL-E mini?
DALL-E mini was published by Craiyon, based in United States of America, 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.