Liquid AI Releases LFM2.5-Encoder-230M and LFM2.5-Encoder-350M: Bidirectional Encoders That Stay Fast at 8K Context on CPU

Short summary

Liquid AI released two open-weight bidirectional encoders, LFM2.5-Encoder-230M and LFM2.5-Encoder-350M, both with 8,192-token context built on the LFM2 hybrid backbone. The 350M model ranks fourth of 14 on a 17-task GLUE/SuperGLUE/multilingual suite, trailing only larger models. The 230M completes one 8K-token forward pass on CPU in about 28 seconds.

• •Two open-weight bidirectional encoders (230M and 350M) with 8K context released
• •350M ranks 4th of 14 models on GLUE/SuperGLUE/multilingual benchmark
• •230M completes an 8K-token CPU forward pass in ~28 seconds

Generated with AI, which can make mistakes.