FLUX.2 [klein] 4B β€” pinned mirror

This is not a Beenga model. It is a byte-for-byte copy of black-forest-labs/FLUX.2-klein-4B at revision e7b7dc27f91deacad38e78976d1f2b499d76a294, kept so that Beenga Image builds stay reproducible and can be rebuilt if the upstream repository is moved, gated or withdrawn.

All credit for these weights belongs to Black Forest Labs. Nothing here is modified, retrained or fine-tuned. If you want this model, prefer the upstream repository β€” it is the source of truth and will carry any corrections.

Why a mirror exists

Every measurement in Beenga Image's model card was made against this exact revision. Pinning the revision makes a rebuild reproducible; mirroring it makes a rebuild possible independent of upstream availability.

Licence

Apache 2.0, as published by Black Forest Labs. LICENSE.md is included in this repository as part of the upstream copy. Use may also be subject to Black Forest Labs' applicable usage policies.

Note the variant. BFL publishes the 4B models under Apache 2.0 β€” including 4b-fp8 and 4b-nvfp4 β€” and the 9B models under the FLUX Non-Commercial License v2.1, with 9b-fp8 additionally gated. The names differ by two characters. Check the licence of any model you substitute.

Downloads last month
36
Safetensors
Model size
4B params
Tensor type
BF16
Β·
Inference Providers NEW
This model isn't deployed by any Inference Provider. πŸ™‹ Ask for provider support

Model tree for beenga8/flux2-klein-4b-mirror

Finetuned
(49)
this model