AIMarkTechPostabout 1 hour ago
Liquid AI DSpark Drafts Speed Decoding 3.18x
9 min read
Liquid AI released DSpark draft checkpoints for LFM2.5-1.2B-Instruct, LFM2.5-2.6B, and LFM2.5-8B-A1B. A 300M-parameter draft proposes nine tokens verified in one pass, delivering up to 3.18x faster decoding on H100 and 2.87x on M4 Max MacBook Pro with identical outputs.
Level
Hype check
Tap to vote and see what everyone thinks.
Summary by ByteBrief
