Liquid AI releases LFM2.5-8B-A1B

In three linesLiquid AI releases LFM2.5-8B-A1B, 8B model with 128K context window, 38T pre-training tokens, and large-scale RL. Doubled vocabulary for non-Latin languages. Supports tool chaining and complex tasks on entry-level laptops.

Summary generated by Claude — human-verified