Liquid AI Launches LFM2.5-2.6B: On-Device Agentic Model Outperforming Larger Counterparts
maximelabonne · x · 2026-08-04
Liquid AI has released LFM2.5-2.6B, now available on Hugging Face. Designed as an on-device agentic model, it plans, calls tools, and handles multi-step tasks entirely locally on phones, laptops, and robots, ensuring data privacy with near-zero marginal cost per run.
Built on a hybrid architecture pre-trained on 34T tokens, the model supports a 128K context length. The team highlights its balanced intelligence per watt and customizability on a single GPU. It demonstrates impressive performance, scoring 77.83 on ToolSandbox (beating Qwen3.5-9B) and outperforming models nearly 4x its size on benchmarks like Multi-IF and IFStruct.
Related event: Liquid AI Releases LFM2.5-2.6B Edge Agent Model(4 posts)→
More from coding & agent
- Developer Shares 7 Tips for Working with 5 Coding Agents Simultaneously — rseroter · 2026-08-04
- Running Codex on 128-Core CPU Clusters: A New Compute Approach — BenBajarin · 2026-08-04
- srt2speech: Open-Source Lightweight SRT Dubbing Tool with Voice Cloning — Charming-Author4877 · 2026-08-04
- Running Local Background Agents at Zero Marginal Cost: Burning >1M Tokens/hr — maximelabonne · 2026-08-04
- Building a Zero-Human Game Studio with Claude Agents — AceHighness · 2026-08-04
- npm Supply Chain Attack Hits 868+ Packages With Credential-Stealing Worm — RSync25 · 2026-08-04