Liquid AI Launches LEAP: A New Platform for On-Device AI Development

Liquid AI Introduces LEAP: A New Frontier in Edge AI Development

Liquid AI, a startup founded by former MIT researchers, has unveiled LEAP (Liquid Edge AI Platform), a revolutionary software development kit (SDK) designed to simplify the integration of small language models (SLMs) into mobile applications 1

1

. This launch marks a significant step towards making on-device AI more accessible and efficient, addressing growing concerns over cloud-based AI services.

Key Features of LEAP

LEAP is engineered to cater to developers with varying levels of machine learning expertise. The SDK can be integrated into iOS or Android projects with minimal code, making the process of calling a local model as straightforward as interacting with a cloud API 1

1

.

Some notable features include:

1. Cross-platform compatibility (iOS and Android)
2. A built-in model library with compact models as small as 300MB
3. Support for both Liquid AI's proprietary Liquid Foundation Models (LFMs) and popular open-source small models
4. Tools for local inference, memory optimization, and device compatibility

Apollo: A Companion App for Model Testing

Alongside LEAP, Liquid AI has introduced Apollo, a free iOS app that allows developers and users to interact with LEAP-compatible models in an offline, local environment 1

1

. Originally a separate startup acquisition, Apollo has been redesigned to support the entire LEAP model library, offering a "TestFlight-like" experience for local AI models 2

2

.

The LFM2 Model Family: Powering LEAP

The LEAP SDK is built on the foundation of Liquid AI's recently announced LFM2 model family. These models, designed specifically for on-device workloads, come in sizes ranging from 350M to 1.2B parameters 1

1

. The LFM2 models are optimized for efficient inference on various hardware types, including CPUs, GPUs, and NPUs 2

2

.

Addressing Industry Needs

LEAP and Apollo arrive at a crucial time when developers are actively seeking alternatives to cloud-only AI services. The platform addresses several key concerns:

1. Latency: By running models locally, LEAP reduces the delay associated with cloud-based processing
2. Cost: On-device processing can be more cost-effective than relying on cloud infrastructure
3. Privacy: Local execution ensures user data remains on the device
4. Offline availability: Apps can function without an internet connection

Accessibility and Licensing

LEAP is currently available for free under a developer license, which includes access to the core SDK and model library. Liquid AI has indicated that premium enterprise features will be offered under a separate commercial license in the future 1

1

.

The LFM2 models are free for academic use and commercial use by companies with under $10 million in revenue. Larger organizations are required to contact Liquid AI for licensing details 1

1

.

Industry Impact and Future Prospects

The introduction of LEAP and Apollo represents a significant shift in the AI development landscape. By making on-device AI more accessible, Liquid AI is paving the way for a new generation of privacy-preserving, efficient AI applications 2

2

.

As the demand for edge AI solutions continues to grow, platforms like LEAP are likely to play a crucial role in shaping the future of mobile and IoT applications. The emphasis on small, efficient models that can run on consumer hardware opens up new possibilities for AI integration across a wide range of devices and use cases.