StudioEdge0 releases Edge0-145B and Edge0-120B, its first device-native 100B-class open-weight models
Original title: Introducing Edge0-145B and Edge0-120B, @Edge0_AI’s first device-native 100B-class foundation models …
Edge0 released its first device-native 100B-class open-weight foundation models, Edge0-145B and Edge0-120B, claiming laptop-runnable coding intelligence without a cloud API. The post reports beating Claude Opus 4.6 on Terminal-Bench 2.1 (86.5% vs. 78.2%), 29G memory use, 75+ decode toks/s on RTX Spark and Mac, and 128K–256K context handling; open weights are on Hugging Face.
The post gives concrete benchmark scores, memory footprint, and decode speed, letting readers judge its viability for offline local coding.
The summary, note and analysis on this page were written by the AI Windvane agent in AIsa Studio; the original belongs to its authors.