Tier 2 Restricted open-weight
What the publisher claims: open-weight, open-weights
The one-liner: The open-weight image family whose own release table shows three different licenses in one row.
The six questions
Weights for FLUX.1 [schnell] and FLUX.1 [dev] are downloadable from the publisher's own repositories; only [pro] is closed.
The fastest member is Apache-2.0, but the stronger [dev] weights sit under a custom non-commercial license and [pro] is API-only, so a family-level permission grant does not exist.
Schnell and dev run locally on consumer GPUs through Diffusers, ComfyUI, and the publisher's inference repo.
Model cards and blog posts describe architecture and evaluation, but the training corpus and recipe are not published.
Limited: a skilled group could not rebuild an equivalent FLUX.1 model from published artifacts.
black-forest-labs repos accept community PRs and discussions; release direction is fully BFL's.
License
Mixed by variant: FLUX.1 [schnell] Apache-2.0; FLUX.1 [dev] FLUX.1-dev Non-Commercial License; FLUX.1 [pro] proprietary API only — license source ↗
- The official model table lists different licenses per variant, so the family label hides the terms of the variant you actually download
- FLUX.1 [dev] weights are released under a non-commercial license
- FLUX.1 [pro] has no downloadable weights at all
What is actually released
- weights (schnell, dev)
- inference code and reference implementation
- model cards
- quantized community variants
Why it matters
A single name can carry Apache-2.0, a non-commercial custom license, and no weights at all — so openness has to be checked per variant, not per family.
Sources
Derived from the openness dataset behind But Is It Open? — snapshot 2026-11-07 ↗. Scores are qualitative readings of the cited evidence, not verdicts; tier measures openness, not intelligence.