Hitchhikethe internet
Directory/Model providers/Liquid AI
Model providers · Tool

Liquid AI

Researched

Liquid AI is an efficiency-first foundation model company building device-native foundation models. It offers open-weight LFMs for on-device deployment across mobile, automotive, and edge applications via the LEAP platform and Liquid Apollo.

Cambridge, MA Checked XLinkedInYouTube
Official websiteVisit website
Official site snapshots

See the official site at a glance

Read-only public captures of Liquid AI’s homepage and verified pricing page. Screenshots are dated, never live embeds, and open full-screen.

Visit live site
Homepage · captured Jul 23, 2026
Pricing page · captured Jul 23, 2026
At a glance

In one minute

Start here for the decision-making essentials: what Liquid AI does, who it is for, how it is accessed, and the first-party sources behind this profile.

Pricing2 options

Custom pricing based on use case is offered for businesses generating more than $10M per y

Exemptions are offered for organizations that do not exceed $10M in annual revenue each ye

Platforms
Liquid Edge AI Platform (LEAP)Windows, macOS, Linux
API accessNot public
FoundedNot disclosed by source
AvailabilityWeb / remote
LicenseEnterprises can license or purchase full local access to LFMs from Liquid AI's library.

Best suited to

Source-backed fit
Developers building on-device AI applications for mobile and edge Enterprises needing privacy-preserving local AI inference Teams deploying small language models in iOS or Android apps Automotive companies embedding in-car intelligence Organizations requiring local fine-tuning of foundation models developers
Decision support

Common questions and adoption checks

6 sourced answers

Short answers to the questions buyers and builders commonly ask about Liquid AI. Each answer cites the shared ledger below, where every source is listed once.

01What does Liquid AI say it can do?

Supports on-device reasoning under 1GB. · deploy small language models in mobile apps with a few lines of code · Customer-ready in-vehicle experiences spanning speech, language understanding, and reasoni · Low-latency, high-efficiency models designed specifically for on-device deployment, suppor

On-device reasoning under 1GB
Ledger citations[1] liquid.ai[2] liquid.ai/blog/liquid-ai-launches-leap-and-apollo-bringing-edge-ai-to-every-developer[3] liquid.ai/press/liquid-ai-and-mercedes-benz-partner-to-scale-embedded-in-car-intelligence[4] liquid.ai/news
02Who is Liquid AI intended for?

developers · Mercedes-Benz customers in North America. · Automotive, consumer electronics, e-commerce, financial services, healthcare, industrial,

developer-first platform for AI on the edge
Ledger citations[2] liquid.ai/blog/liquid-ai-launches-leap-and-apollo-bringing-edge-ai-to-every-developer[3] liquid.ai/press/liquid-ai-and-mercedes-benz-partner-to-scale-embedded-in-car-intelligence[4] liquid.ai/news
03What use cases does Liquid AI describe?

on-device AI deployment for mobile apps · LEAP enables edge AI applications, covering model selection, inference, customization, eva · Embedded in-car intelligence: voice control, vehicle functionality, and contextual underst · Embedded in-car intelligence for automotive (Mercedes-Benz partnership)

developer-ready platform for on-device AI deployment
Ledger citations[2] liquid.ai/blog/liquid-ai-launches-leap-and-apollo-bringing-edge-ai-to-every-developer[6] liquid.ai/faq[3] liquid.ai/press/liquid-ai-and-mercedes-benz-partner-to-scale-embedded-in-car-intelligence[4] liquid.ai/news
04What should teams verify before adopting Liquid AI?

requires devices with 4GB+ RAM

Runs smoothly on devices with 4GB+ RAM
Ledger citation[2] liquid.ai/blog/liquid-ai-launches-leap-and-apollo-bringing-edge-ai-to-every-developer
05What pricing information is available for Liquid AI?

Custom pricing based on use case is offered for businesses generating more than $10M per y · Exemptions are offered for organizations that do not exceed $10M in annual revenue each ye

We offer custom pricing based on use case for businesses generating more than $10M per year in annual revenue.
Ledger citation[6] liquid.ai/faq
06Does Liquid AI document API access?

Software development kits (SDKs) available for download

sign up for or download our Services, including any of our software development kits
Ledger citation[9] liquid.ai/privacy-policy
Decision guide

Capabilities and operating fit

Model providers

This profile connects the jobs Liquid AI is described as handling with its delivery model, access options and the subjects used to match it to related products in this directory.

Common use cases

  • On-device AI deployment for mobile apps
  • Edge AI applications spanning model selection, inference, customization, evaluation, and o
  • Embedded in-car intelligence via Mercedes-Benz partnership
  • Drug discovery foundation models via Insilico Medicine partnership
  • Structured visual intelligence from edge to cloud
  • on-device AI deployment for mobile apps

Access signals

Pricing model
Custom pricing based on use case is offered for businesses generating more than $10M per y · Exemptions are offered for organizations that do not exceed $10M in annual revenue each ye
API
Not publicly listed
Source links
16 recorded
Source-backed

Verified facts

Updated July 23, 2026

Each fact points to a recorded source, making it easy to distinguish verified product information from claims that need checking.

Official website

HTTP 200 verified twice

First-party description

Liquid AI — Device-native foundation models.

Source-supported facts

Efficiency-first foundation model company · Device-native foundation models · Deployed on phones, laptops, cars, space, e-commerce, financial services, bio, defence

Company

Liquid AI is an efficiency-first foundation model company building device-native foundation models.

Deployment

Models are deployed on phones, laptops, cars, space, e-commerce, financial services, bio, and defence.

Platform

Runs on llama.cpp, MLX, ONNX, CoreML, SGLang, and vLLM runtimes.

Security

Designed so that user data never leaves the device.

Capability

Supports on-device reasoning under 1GB.

View 32 more verified facts
Fine tuning

Users can fine-tune an LFM to their own data.

Model

LFM2.5-VL-450M is shipping for structured visual intelligence, edge to cloud.

Model

LFM2.5-350M was trained on 28T tokens and targets on-device use.

Model

LFM2-24B-A2B is an MoE model scaling up the LFM2 architecture.

Model

LFM2.5-8B-A1B is an on-device Mixture of Experts model.

Integration

Has a partnership with Mercedes-Benz to scale embedded in-car intelligence.

Integration

Has a strategic partnership with Insilico Medicine for drug discovery foundation models.

Release date

2025-07-15

[2]liquid.ai/blog/liquid-ai-launches-leap-and-apollo-bringing-edge-ai-to-every-developer
Model family

LFM2

[2]liquid.ai/blog/liquid-ai-launches-leap-and-apollo-bringing-edge-ai-to-every-developer
Platform

Android; iOS

[2]liquid.ai/blog/liquid-ai-launches-leap-and-apollo-bringing-edge-ai-to-every-developer
Deployment

on-device (edge)

[2]liquid.ai/blog/liquid-ai-launches-leap-and-apollo-bringing-edge-ai-to-every-developer
Capability

deploy small language models in mobile apps with a few lines of code

[2]liquid.ai/blog/liquid-ai-launches-leap-and-apollo-bringing-edge-ai-to-every-developer
Application type

iOS-native mobile application

[2]liquid.ai/blog/liquid-ai-launches-leap-and-apollo-bringing-edge-ai-to-every-developer
Use case

on-device AI deployment for mobile apps

[2]liquid.ai/blog/liquid-ai-launches-leap-and-apollo-bringing-edge-ai-to-every-developer
Audience

developers

[2]liquid.ai/blog/liquid-ai-launches-leap-and-apollo-bringing-edge-ai-to-every-developer
Security

fully private, 100% local, privacy-preserving

[2]liquid.ai/blog/liquid-ai-launches-leap-and-apollo-bringing-edge-ai-to-every-developer
Limitation

requires devices with 4GB+ RAM

[2]liquid.ai/blog/liquid-ai-launches-leap-and-apollo-bringing-edge-ai-to-every-developer
Distribution

Google Play; App Store

[2]liquid.ai/blog/liquid-ai-launches-leap-and-apollo-bringing-edge-ai-to-every-developer
Distribution

Certain models are available on the playground (free with rate limits) and on Openrouter (paid with higher limits).

[6]liquid.ai/faq
Distribution

All models are available for direct download on Hugging Face.

[6]liquid.ai/faq
Platform

The majority of models are offered for customization and deployment through the LEAP platform.

[6]liquid.ai/faq
Pricing

Custom pricing based on use case is offered for businesses generating more than $10M per year in annual revenue.

[6]liquid.ai/faq
Pricing

Exemptions are offered for organizations that do not exceed $10M in annual revenue each year.

[6]liquid.ai/faq
Use case

LEAP enables edge AI applications, covering model selection, inference, customization, evaluation, and on-device testing.

[6]liquid.ai/faq
Fine tuning

LFMs can be fine-tuned; an on-prem LFM customization stack is available for purchase to enterprises.

[6]liquid.ai/faq
License

Enterprises can license or purchase full local access to LFMs from Liquid AI's library.

[6]liquid.ai/faq
Deployment

LFMs come with two software stacks: the LFM inference stack and the LFM customization stack.

[6]liquid.ai/faq
Open source

Models are open weight (weights publicly available for download, local deployment, and commercial use), but not open source; training code, data, and methodology are not released.

[6]liquid.ai/faq
Company

Liquid AI

[9]liquid.ai/privacy-policy
Platform

Web-based Services and downloadable Application

[9]liquid.ai/privacy-policy
Deployment

Cloud-hosted services accessible via website

[9]liquid.ai/privacy-policy
Api

Software development kits (SDKs) available for download

[9]liquid.ai/privacy-policy
Practical capabilities

What it helps with

8 documented areas

A concise view of the jobs, capabilities and integrations described in the recorded product sources.

Use case

On-device AI deployment for mobile apps

Use case

Edge AI applications spanning model selection, inference, customization, evaluation, and o

Use case

Embedded in-car intelligence via Mercedes-Benz partnership

Use case

Drug discovery foundation models via Insilico Medicine partnership

Use case

Structured visual intelligence from edge to cloud

Use case

LEAP enables edge AI applications, covering model selection, inference, customization, eva

Use case

Embedded in-car intelligence: voice control, vehicle functionality, and contextual underst

Use case

Embedded in-car intelligence for automotive (Mercedes-Benz partnership)

Availability

Where it runs and where to get it

Source checked

Documented product formats, platforms and official distribution destinations. Availability can vary by region and plan.

Cost / license

Custom pricing based on use case is offered for businesses generating more than $10M per y · Exemptions are offered for organizations that do not exceed $10M in annual revenue each yeEnterprises can license or purchase full local access to LFMs from Liquid AI's library.Models are open weight (weights publicly available for download, local deployment, and comEnterprises can license or purchase full local access to LFMs from Liquid AI's library. · Models are open weight (weights publicly available for download, local deployment, and com

Application types

iOS-native mobile applicationOn-device mixture-of-experts foundation modelAutomated Foundation Model Design (Liquid Labs)

Origin

MIT-born

Platforms

Runs on llama.cpp, MLX, ONNX, CoreML, SGLang, and vLLM runtimes.Android; iOSThe majority of models are offered for customization and deployment through the LEAP platfWeb-based Services and downloadable ApplicationTarget platform is Mercedes-Benz third- and fourth-generation MBUX, built on the in-house Liquid Edge AI Platform (LEAP) for on-device AI deploymentLiquid Edge AI Platform (LEAP)AMD Ryzen and Ryzen AI processorsWindows, macOS, Linux

App stores and other links

GitHubGoogle PlayApple App StoreGitHub
Implementation details

Adoption notes

DeploymentModels are deployed on phones, laptops, cars, space, e-commerce, financial services, bio, · on-device (edge) · LFMs come with two software stacks: the LFM inference stack and the LFM customization stac · Cloud-hosted services accessible via website
LicenseEnterprises can license or purchase full local access to LFMs from Liquid AI's library. · Models are open weight (weights publicly available for download, local deployment, and com
Model supportLFM2.5-VL-450M is shipping for structured visual intelligence, edge to cloud. · LFM2.5-350M was trained on 28T tokens and targets on-device use. · LFM2-24B-A2B is an MoE model scaling up the LFM2 architecture. · LFM2.5-8B-A1B is an on-device Mixture of Experts model. · Liquid Foundation Models (LFMs)
Data controlDesigned so that user data never leaves the device. · fully private, 100% local, privacy-preserving · Privacy policy addresses data security and personal data protection · Privacy-preserving on-device speech processing with no continuous data exchange with the c · Fully private intelligence on device
Learning curveIntermediate
Primary use casesOn-device AI deployment for mobile apps, Edge AI applications spanning model selection, inference, customization, evaluation, and o, Embedded in-car intelligence via Mercedes-Benz partnership, Drug discovery foundation models via Insilico Medicine partnership, Structured visual intelligence from edge to cloud, on-device AI deployment for mobile apps, LEAP enables edge AI applications, covering model selection, inference, customization, eva, Embedded in-car intelligence: voice control, vehicle functionality, and contextual underst, Embedded in-car intelligence for automotive (Mercedes-Benz partnership), Automotive, Healthcare, Financial services, Defense, Industry solutions across Automotive, Consumer electronics, E-commerce, Financial services, On-device AI solutions for consumer laptops and enterprise endpoints, AI glasses / smart wearables (scene understanding)

What to verify before adopting

  • requires devices with 4GB+ RAM
Evolution and major updates

Liquid AI timeline

A concise history of software releases and material product changes. Events appear only when a dated source supports what changed.

Research in progress
Scheduled for research

Building a reliable release history.

This profile is being checked for dated releases and material product changes. Nothing appears here until the exact date and event can be verified from a recorded source.

Dated event Short explanation Original source
Citation ledger

Recorded sources

16 unique pages

Facts, answers, structured details, milestones and primary resource links cite this shared ledger. Each external page appears once; release tags from the same GitHub project are grouped under one release history.

  1. 1liquid.ai 17 facts · 1 answer · Official site
  2. 2liquid.ai/blog/liquid-ai-launches-leap-and-apollo-bringing-edge-ai-to-every-developer 11 facts · 4 answers · Official site
  3. 3liquid.ai/press/liquid-ai-and-mercedes-benz-partner-to-scale-embedded-in-car-intelligence 12 facts · 3 answers · Official site
  4. 4liquid.ai/news 10 facts · 3 answers · Official site
  5. 5liquid.ai/blog/liquids-edge-ai-platform-leap-expands-support-to-laptops-with-best-in-class-performance-on-amd-ryzen-tm-and-ryzen-ai-tm-processors 12 facts · Official site
  6. 6liquid.ai/faq 10 facts · 2 answers · Official site
  7. 7liquid.ai/case-studies 10 facts · Official site
  8. 8liquid.ai/news/models 10 facts · Official site
  9. 9liquid.ai/privacy-policy 6 facts · 1 answer · Security
  10. 10liquid.ai/press/liquids-edge-ai-platform-leap-expands-support-to-laptops-with-best-in-class-performance-on-amd-ryzen-tm-and-ryzen-ai-tm-processors 6 facts · Official site
  11. 11liquid.ai/models 1 fact · Official site
  12. 12liquid.ai/news/case-studies Official site
  13. 13liquid.ai/news/company-news Official site
  14. 14liquid.ai/news/research Official site
  15. 15liquid.ai/press/brilliant-labs-partners-with-liquid-ai-to-bring-vision-language-tech-to-your-glasses Official site
  16. 16liquid.ai/press/g42-and-liquid-ai-partner-to-deliver-private-local-and-efficient-ai-solutions-to-enterprises-at-scale Official site
Research status120 substantive facts · 16 source pages · quality score 100/100