Liquid AI unveils on-device memory layer and LFM2.5-2.6B embedded agent at Snapdragon Summit
samcharrington · x · 2026-09-24
At the Snapdragon Summit, Liquid AI CEO Ramin Hasani announced Liquid Context on Snapdragon, a memory layer for on-device agents with read, compress, and write capabilities.
Key points:
- With user permission, device signals are turned into an understanding of routines, preferences, and needs
- That personal context serves any agent the user chooses — local, cloud, or hybrid
- A companion Liquid Agent runs on LFM2.5-2.6B, optimized for Qualcomm's Hexagon NPU
- OEMs get a shared foundation for personal context plus a ready-built embedded agent
The announcement followed a day-2 fireside chat between Qualcomm CEO Cristiano Amon and Hasani on co-designing hardware, models, harnesses, and agents for efficient on-device AI at scale.
Related event: Liquid AI Brings On-Device Context Layer to Snapdragon NPUs with Qualcomm(3 posts)→
More from coding & agent
- 30 lines of JavaScript, no image models: pure-code generative art demo — nc_frey · 2026-09-24
- Developer calls Claude Opus 5.5 'a doof' at database tasks — rickasaurus · 2026-09-24
- TRACES: A New Benchmark That Grades AI Problem-Solving Process, Not Just Correct Answers — dr_cintas · 2026-09-24
- Dev reverse-engineers Qwen Image 2.1 PE, ships ComfyUI node that auto-computes dimensions — BleynSpecnaz · 2026-09-24
- Cua Releases Cua-S1-4B, First Multimodal Decision Model RL-Trained on Live Computer-Use Tasks — multimodalart · 2026-09-24
- Applied Compute uses Jev to auto-cluster failure modes in RL training traces — rhythmrg · 2026-09-24