LFM2.5-DSpark: Up to 3.2x Faster Inference from H100 to MacBook
Liquid AI released DSpark draft-model checkpoints for three models in its LFM2.5 family, adding a speculative decoding path that delivers up to 3.18x faster decoding on GPU and up to 2.87x on-device without changing output quality. It is the company's first public release of speculative decoding models for Liquid Foundation Models.
Aug 20, 2026
Liquid AI
·
Company Website