LFM2.5-230M

Name: LFM2.5-230M
Author: Liquid Ai

Liquid Ai🇺🇸 United States

active

Compare with other models →

Context window33K tokens

Version History

2.5minorJune 26, 2026

LFM2.5-230M is a new compact model in the LFM family, built on LFM2 architecture with extended pre-training and reinforcement learning. It features 230M parameters, tool use capabilities through distillation from LFM2.5-350M, and optimized edge inference performance.

Benchmark Scores

Full leaderboard →

25.4%

GPQA

213.0 tokens_per_sec

Speed (tok/s)

Coverage

model releaseLiquid Ai

Liquid AI releases LFM2.5-230M, a 230M parameter edge model running at 213 tok/s on Galaxy S25 Ultra

Liquid AI has released LFM2.5-230M, a 230M parameter hybrid model trained on 19 trillion tokens with a 32,768 token context window. The model achieves 213 tok/s decode speed on Galaxy S25 Ultra and 42 tok/s on Raspberry Pi 5, with support for function calling and data extraction tasks.

June 26, 2026 · 12:21 AM2 min read

Liquid AI LFM2.5 edge AI