LFM3-1.7B beats its 2.6B predecessor with 35% fewer params, 256K context
JosephJacks_ · x · 2026-10-28
Dev javierbits highlights the new open-weight LFM3 family from Liquid AI: LFM3-1.7B beats the previous 2.6B model with 35% fewer parameters, offers a 256K context window, and is 1 GB at 4-bit, fine-tunable on a single GPU. The lineup spans text, VL (video input), audio (tool calls, 18 languages), omni, plus d1 decision models with typed outputs and calibrated probabilities in a single pass.
More from Multimodal
- Google researcher uses Genie 4 to revisit old homes and talk to his younger self — holynski_ · 2026-10-28
- Open-source pipeline animates non-humanoid 3D assets with video gen and DINOv2 joint tracking — mishig25 · 2026-10-28
- Fixing AI image drift: use the approved design as a reference to remove noise — LudovicCreator · 2026-10-28
- One bag, three brand identities: comparing AI variants on the same board — LudovicCreator · 2026-10-28
- One Firefly Board, one coffee brand: full identity workflow from moodboard to launch poster — LudovicCreator · 2026-10-28
- MiniMax H3 generates stunningly realistic in-car selfie, user calls it best work yet — Error4010 · 2026-10-28