What Liquid AI does
Liquid AI builds device-native foundation models, called LFMs, designed to run on phones, laptops, cars and other hardware outside the data center within tight latency, privacy and hardware limits. It is aimed at developers building local or edge AI and at enterprises in sectors such as automotive, consumer electronics, financial services, healthcare and industrial robotics. The site names open-weight models, decision models (d1), vision-language models and encoders, fine-tuning, and deployment to runtimes including llama.cpp, MLX, ONNX, CoreML and vLLM. Pricing combines a free open-weight license for businesses under $10M in annual revenue, an enterprise license, per-token API pricing such as $0.05 per million input tokens for LFM2.5-24B-A2B, and custom deployments through sales.
How Liquid AI compares with its closest alternatives
3 of the 5 closest alternatives have a free plan, as Liquid AI does. All 5 offer an API, as Liquid AI does. Liquid AI was founded in 2023; the alternatives were founded between 2020 and 2024, 1 of them earlier. Liquid AI is based in Boston, United States; none of its 5 alternatives is in the United States. They are in France, China, the United Arab Emirates and elsewhere.