Liquid AI released LFM2.5-8B: compact MoE model with 128K context
Liquid AI released LFM2.5-8B-A1B — a MoE model that activates only 1.5B of 8.3B parameters. Runs on standard PCs, supports 128K token context, capable of reason
◐ Listen to article
Liquid AI released LFM2.5-8B-A1B — a MoE model that activates only 1.5B of 8.3B parameters. Runs on standard PCs, supports 128K token context, capable of reasoning and tool calling. Ideal for local AI applications.