FLUX.1 [schnell]

Open-weight image generation and editing alternative; FLUX.1 schnell uses Apache 2.0, distinct from FLUX dev licensing.

Score calculation

(4 + 1 + 8 + 8 + 1 + 7) ÷ 6 = 4.8

6 of six dimensions rated. Unrated dimensions are excluded from the average.

Setup details

FLUX.1 [schnell]: both access. Open weights under Apache 2.0 for local use; API pricing is usage-based. Ratings describe the stated product, company or configuration in each finding.

What fits your needs

FLUX.1 [schnell] open-weight image model; this scope excludes FLUX dev/pro models, which use different licenses and routes.

Tasks

Text-to-image generation can run using local inference code or through an API provider; the schnell variant is distributed under Apache 2.0.

See the evidence

Billing and setup

Local model weights have no per-generation provider fee, but require compatible hardware and setup; hosted API usage is charged by provider/usage.

See the evidence

Tradeoffs to expect

Local requirements

Local inference requires a GPU/runtime with sufficient memory; exact performance and workable quantization vary by hardware and implementation, and no single consumer minimum is asserted.

See the evidence

Route

A hosted API or community interface may use a different provider and privacy terms than local inference; check the endpoint before uploading references.

See the evidence

How to switch

Get started

Download the exact schnell weights and inference code from the official model sources for local use, or choose an API provider and review its current rate and terms.

See the evidence

Practical questions

Age and sensitive-use rules

The FLUX.1 schnell model card and license identify weights and terms but do not set a consumer minimum age or hosted-service safety policy; those depend on the chosen provider and local operator.

See the evidence

Likeness and consent

The reviewed model card and Apache license do not specify consent checks, NCII moderation or impersonation filters. A local runtime has no verified provider-side enforcement; check the selected interface separately.

See the evidence

Labels and corrections

The model card does not specify content credentials, generated-image labels or a correction/report mechanism for local use. These controls depend on the inference application.

See the evidence

FLUX.1-schnell model card — Black Forest Labs on Hugging Face ↗

Energy practices

4/10↑

Black Forest Labs publishes a model card with local deployment instructions and four-step inference parameters. No serving-route energy or water measurements or clean-power sourcing were found.

See the evidence

Black Forest Labs publishes a model card with local deployment instructions and four-step inference parameters. No serving-route energy or water measurements or clean-power sourcing were found. The score describes documented measurement transparency and available controls, not environmental performance.

1 + 1 reporting + 0 electricity + 2 controls = 4/10. This grade covers documented practices and transparency; it does not measure the energy or emissions of a request.

The model can run on user hardware or through separately operated hosted providers; electricity and water vary by hardware, provider, region, and implementation. Four steps does not quantify joules per image or quality-adjusted impact.

Reporting1/3: The official model card documents model architecture/use and local runtime, but no quantitative energy/carbon/water report or facility boundary.
Electricity sourcing0/3: FLUX.1 [schnell] is available for local use and through hosted API/provider paths; the chosen route determines electricity sourcing and none is declared for this assessed setup.
Resource controls2/3: The inference code exposes step count and output dimensions, and the model card demonstrates four-step generation. This is a practical set of controls, but no physical-resource saving has been measured for those settings.

Water disclosure

Water reporting has not been assessed for this choice.

Training data

1/10↑

The model card says FLUX.1 [schnell] was distilled from a larger model and provides an Apache 2.0 reuse license; it does not disclose training sources or a training-data rights basis.

See the evidence

The identifiable release is assessable, but the published card describes model reuse and distillation rather than the data rights for its base or inherited training stages. Open weights and code don’t settle training permissions.

1 + 0 source transparency + 0 permission + 0 creator controls = 1/10. This grades documented training-data safeguards, not a legal ruling.

The training data, FLUX.1 [dev] source rights, distillation data and creator controls are undisclosed.

Source transparency0/3: The release card identifies distillation but provides no meaningful training-source account.
Documented permission0/4: An Apache or model-use license governs reuse of weights, not permission for the training data.
Creator controls0/2: No relevant creator opt-out, removal or permission control is documented in the reviewed model cards.

Data-rights rubric ↗

Your data & privacy

8/10↑

A fully local offline deployment can keep prompts and images on the operator’s device; the model card also lists hosted API routes with different data boundaries.

See the evidence

This score applies only to local offline inference. The official model card documents local Diffusers and ComfyUI paths as well as hosted API sources; the model itself doesn’t confirm privacy terms for those hosted routes.

Openness

8/10↑

FLUX.1 schnell is available as an Apache 2.0 checkpoint, with reference inference code and instructions for local Diffusers/ComfyUI use; this does not reveal the training data.

See the evidence

The official Black Forest Labs model repository identifies the exact schnell checkpoint, Apache 2.0 license, inference code, and local use paths. This rating applies to schnell only and does not generalize to dev variants with different terms.

Worker protections

1/10↑

See the evidence

The FLUX.1 schnell model card documents a permissively licensed downloadable release and provides model-use context, but does not report workforce or supplier conditions, pay, worker voice, or independent data-worker outcomes. Its open license also means downstream use is not evidence of Black Forest Labs’ own deployment or contracts.

1 + 0 pay & conditions + 0 worker voice & remedy + 0 evidence & follow-through = 1/10. This grades documented worker safeguards, not a measured rate of mistreatment.

Worker pay, exposure and health protections, paid versus volunteer roles, vendor/subcontractor coverage, grievance routes, and independent outcome checks are not established by the reviewed one official model card and release license surface checked; company-wide workforce, other model families, and data suppliers are not covered by that card.

Pay & conditions0/3: No opened source reports worker pay, hours, exposure protections, contractor coverage, or worker-level outcomes for Black Forest Labs. The reviewed material is limited to One official model card and release license surface checked; company-wide workforce, other model families, and data suppliers are not covered by that card.; its silence is an information gap, not evidence of poor conditions.
Worker voice & remedy0/3: The reviewed material for Black Forest Labs doesn’t show worker representation, grievance access, bargaining voice, or documented remedies. Public code contributions, customer controls, or creator rights do not demonstrate worker voice across employees and suppliers.
Evidence & follow-through0/3: No reviewed source provides an independent audit or corroborated worker-level outcomes for Black Forest Labs’s relevant employees, contractors, and data-work suppliers. The available disclosures are company/project materials rather than independent labor verification.

Workers rubric ↗

Military ties

7/10Not established↑

BFL’s hosted/API terms and FLUX [dev] license prohibit military and surveillance purposes, but those restrictions do not bind the Apache-2.0 FLUX.1-schnell weights; no BFL military relationship is established.

See the evidence

BFL’s Usage Policy and FLUX [dev] Non-Commercial License expressly prohibit military, surveillance and law-enforcement purposes for hosted use and the [dev] release/derivatives. The FLUX.1-schnell weights are separately distributed under Apache-2.0 terms, which do not impose a use ban; the hosted or [dev] policy cannot be inherited by that permissive release. Accordingly the company-wide restriction score is zero for the permissive assessed release, while other BFL offerings have distinct contractual rules. No military customer or deployment was established.

7 base + 0 safeguard adjustment = 7/10 (2 of 9 safeguard points; every 3 adds 1). No military relationship was established in the cited sources. This ranks above known defense ties without claiming a verified clean record.

Unpublished contracts, reseller/subcontractor activity, customer-configured use and enforcement outcomes remain unknown. Hosted AUPs do not necessarily apply to independent local deployments or permissively licensed copies.

Use restrictions0/4: FLUX.1-schnell is permissively Apache-2.0 licensed and has no enforceable use ban; the hosted AUP and FLUX [dev] non-commercial license separately bar military/surveillance use but cannot be imputed to Schnell copies or unrelated downstream use.
Enforcement & oversight1/3: For hosted/API and FLUX [dev] routes BFL describes content moderation, appeals and suspension/termination; this enforcement does not cover independently distributed Apache-licensed Schnell copies and no independent outcomes are reported.
Relationship disclosure1/2: BFL publicly states the distinct restrictions and release terms, but reviewed sources do not identify defense customers or deployments.

The public materials disclose open FLUX models and hosted offerings but the targeted official review did not establish a defense customer or operational deployment.

Military rubric ↗

BDS targets

Not listed↑

Categorical check of the BNC campaign guide and No Tech campaign page; absence from these named-target pages is not BDS approval.

See the evidence

No named target for this company was found in the two checked BNC sources as of 2026-10-24. This is a narrow list check, not BDS approval, and it does not capture unlisted relationships or partners.

All sources