Liquid AI Launches LFM2.5-24B-A2B-Base Model
Liquid AI introduces the LFM2.5-24B-A2B-Base, a 24B-parameter hybrid model optimized for on-device deployment.
The LFM2.5-24B-A2B-Base from Liquid AI is a new hybrid model designed for on-device deployment, featuring 24 billion parameters with only 2 billion active parameters. This model supports a context window of 131,072 tokens and a vocabulary size of 128,000 tokens, making it suitable for tasks requiring extensive fine-tuning, such as language-specific assistants or domain-specific applications. It is available for deployment via API, cloud, or local runtimes, including formats like ONNX for cross-platform deployment and MLX for Apple Silicon. Liquid AI claims that the model’s pre-training and mid-training phases, which include tool calls and multi-turn trajectories, prepare it for agentic workloads, as stated in their documentation.