Running Liquid AI LFM V2.5 on my 16 GB MacBookPro

Description and use

Liquid AI’s LFM 2.5 3B runs locally on a 16 GB MacBook Pro, producing reasonable text and code outputs in under 2 minutes.

Problem

Running a 3 billion‑parameter language model locally on a 16 GB MacBook Pro.

Outcome

The 3 billion‑parameter LFM 2.5 model runs locally on a 16 GB MacBook Pro, producing reasonable text and code outputs in under 2 minutes, with no hallucinations and good prompt adherence.

Workflow

• 1. Install required dependencies on MacBook Pro
• 2. Load LFM 2.5 3B model
• 3. Prompt model with "what is a dolphin?" for text generation
• 4. Prompt model with a Python utility request for word counting
• 5. Observe model’s web‑search behavior and file creation
• 6. Evaluate output quality and performance

Evidence

We’re testing LFM 2.5 3B, a 3 billion model from Liquid AI

Source

The text generation is very fast and the answer is more than reasonable. No hallucination and good prompt adherence.

Source

Coding is decent. The model spent about 2 minutes on the task and delivered a working Python file.

Source