ApX logoApX logo

FLUX.2 Dev

Parameters

15B

Modality

Image

Auxiliary Parameters

80M

Architecture

Diffusion Transformer

License

FLUX.2-dev Non-Commercial License

Release Date

15 Jan 2025

API Pricing (per 1M)

Self-hosted only

Hardware & System Requirements

Precision

Native

Workload Memory: 10.80 GB

INT4 · 1024x1024

DiT Backbone Weights: 6.98 GB

VAE Decode Workspace: 1.31 GB

Runtime & Graphs: 2.50 GB

Minimum Required Hardware

1x NVIDIA L4 24GB

10.8 / 21.6 GB usable

1x RTX 3060 12GB

10.8 / 10.8 GB usable

Apple M4 24GB

10.8 / 18.0 GB usable

Determine requirements for your specific workload:

Open in VRAM Calculator

Runtime requirements

Checkpoint format

FP16

CUDA

>= 11.8

PyTorch

>= 2.1

Transformers

>= 4.40.1

ArchitecturePrecisionAttention kernel

Blackwell, Hopper

Native FP16

FlashAttention-3

Ada Lovelace, Ampere

Native FP16

FlashAttention-2

Architecture Diagram

About FLUX.2 Dev

FLUX.2 Dev is Black Forest Labs' next-generation 15B parameter diffusion transformer, extending dual attention mechanisms with deeper joint representations for photorealism and typographic accuracy.

Specifications

Attention

Attention Heads

48

Attention Head Dimension

128

Dimensions

Base Precision

FP16

Auxiliary Parameters

80M

Hidden Dimension Size

15,360

Number of Layers

56

About Flux 2


Other Flux 2 Models