Ethical Stack.

Liquid Nanos

Small specialist models for document extraction, search retrieval, classification, translation, maths and meeting summaries.

Score calculation

(4 + 4 + 8 + 6 + 2 + 2) ÷ 6 = 4.3

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

Full scoring method ↗
Setup details

Liquid Nanos: LFM2.5 encoders, Embedding/ColBERT-350M and VL-Extract variants; LFM2 Japanese PII/ENJP translation, Math and Transcript specialists.

What fits your needs

Liquid Nanos: LFM2.5 encoders, Embedding/ColBERT-350M and VL-Extract variants; LFM2 Japanese PII/ENJP translation, Math and Transcript specialists.

What it does

Small specialist models for document extraction, search retrieval, classification, translation, maths and meeting summaries.

See the evidence
Tradeoffs to expect

Hardware and formats

Check the selected checkpoint and runtime; support differs across text, vision, audio and decision models.

See the evidence
How to switch

Get started

Choose a checkpoint from the official library and a compatible runtime; verify the license before commercial use.

See the evidence
Practical questions

Data path

Local mode keeps model computation on your hardware; remote tools or backends can send data elsewhere.

See the evidence

Use limitations

Vision and omni models understand images and video; this is not an image or video generation service. Decision models return structured decisions, not a general chat answer.

See the evidence

Energy practices

4/10↑

Local deployment and smaller models give you useful controls; clean-electricity evidence is missing.

See the evidence

This rating credits documented practices, not assumed energy savings. The model library and local guides document hardware/runtime choices, smaller checkpoints and quantized formats where supported. Actual consumption depends on the chosen model, device and workload.

1 + 1 reporting + 0 electricity + 2 controls = 4/10. This grades documented practices and transparency.

Comparable task-level electricity, water, training footprint and independently verified clean-power sourcing are not published for this choice.

Reporting1/3
The documented route identifies where inference can run, without an audited environmental footprint.
Electricity sourcing0/3
No attributable electricity procurement or hourly clean-power evidence was found in the cited material.
Resource controls2/3
Model/runtime and size choices provide several controls, without measured energy savings.

Water disclosure

Water reporting has not been assessed for this choice.

Training data

4/10↑

Liquid describes broad training sources, but has not published a permission audit for the whole training mix.

See the evidence

The model documentation describes its base model and training approach. Liquid’s privacy policy adds public internet material and business datasets by agreement. That gives some provenance and a limited permission basis, but no itemized evidence clearing all underlying text, images, speech or inherited training data. The model license grants use of weights; it does not settle training-data rights.

1 + 1 source transparency + 1 permission + 1 creator controls = 4/10, with the permission cap applied.

Dataset-level licenses, consent for speech/images, rights checks for inherited training data, and removal outcomes remain unclear.

Source transparency1/3
Broad source classes and model lineage are described; no complete dataset ledger is published.
Documented permission1/4
Business dataset agreements provide a limited stated permission basis; the wider public-web mix has no inspectable permission audit.
Creator controls1/2
The privacy policy offers personal-data objections, including training use where applicable; this is narrower than a general creator opt-out.
Data-rights rubric ↗

Your data & privacy

8/10↑

Running locally can keep prompts on your own hardware; your app and tools still control other data paths.

See the evidence

Downloadable weights can be run without sending prompts to Liquid once setup is complete. The runtime, remote tools, logs and backups remain your responsibility.

8/10 credits a documented local data path with configuration and audit limits.

No independent network audit is claimed. App backups, runtime telemetry, tools and hosting setup can create other copies.

Openness

6/10↑

Weights are downloadable, with commercial use restricted by annual revenue.

See the evidence

Liquid provides weights, model cards and runtime instructions, but uses a custom license. Commercial users at $10M annual revenue or above need a separate license, so these weights do not offer unrestricted Apache-style commercial use. The complete training data is also unavailable.

6/10 credits inspectable downloadable weights and practical deployment, with a commercial revenue limit and incomplete training disclosure.

The full training dataset and reproducible training pipeline are unavailable.

Worker protections

2/10↑

Liquid publishes hiring and wellbeing commitments, but little evidence about its wider AI workforce.

See the evidence

Its careers page commits to equal opportunity and interview accommodations; company values mention wellbeing. These earn limited policy credit. The reviewed sources provide no contractor pay standards, protected grievance channel or independent evidence covering annotation, moderation and other AI supply-chain work. These findings concern the developer or operator. This product’s use in a particular deployment needs separate evidence.

1 + 1 + 0 + 0 = 2/10. Documented worker protections; missing disclosure is an evidence gap.

Contractor conditions, worker representation, grievance outcomes and independent workforce audits remain undisclosed.

Pay & conditions1/3
Public equal-opportunity and accommodation commitments provide basic policy evidence.
Worker voice & remedy0/3
No worker grievance or representation safeguard was found in the reviewed public material.
Evidence & follow-through0/3
No independent worker-level implementation review was supplied in these sources.
Workers rubric ↗

Military ties

2/10Operational military support↑

Liquid actively offers tactical AI for threat detection, ISR and uncrewed systems.

See the evidence

Liquid’s defense page markets on-device models for combat-relevant intelligence and autonomous systems. We treat that operational offering as direct military enablement under the rubric, which caps the score at 2. Its general service terms prohibit unlawful harm and permit suspension, but do not rule out defense use. This is evidence of an offering, not proof that any particular checkpoint has been deployed by a named military customer. These findings concern the developer or operator. This product’s use in a particular deployment needs separate evidence.

1 + 1 use restrictions + 1 enforcement & oversight + 1 relationship disclosure = 4, capped at 2/10. Known defense ties lower this rating; safeguards can improve it within the relationship’s cap.

Named defense contracts, actual deployments, customer screening and independent oversight are not disclosed on the reviewed page.

Use restrictions1/4
General prohibitions cover unlawful harm; no blanket restriction on operational military use is stated.
Enforcement & oversight1/3
Terms allow suspension for violations, without a disclosed military-use review process.
Relationship disclosure1/2
The defense offering is public, but customers, contract scope and deployments are not named.

The vendor explicitly offers tactical threat detection, ISR, dismounted-team decision support and onboard perception for autonomous uncrewed systems; this is operational military enablement, without a claim of a named contract. The relationship limits this grade to 2/10.

Military rubric ↗

BDS targets

Not listed↑

Liquid AI is not named in the checked BNC boycott and technology target lists.

See the evidence

The BNC consumer guide and No Tech campaign page do not name Liquid AI as a boycott or pressure target. This is a limited campaign-list finding; it is not an endorsement or a wider assessment of every supplier or customer. The defense offering is assessed separately in military ties. These findings concern the developer or operator. This product’s use in a particular deployment needs separate evidence.

The checked lists do not identify every commercial relationship or local campaign.

More from Liquid AI

All sources

Side by side