ππ Introducing Qwen3.7-Plus β a multimodal agent model that unifies vision and language into one versatile agent foundation. β Multimodal interactive hybrid agent: unified GUI & CLI operation across visual and text tasks β Versatile coding agent & productivity assistant with https://t.co/T3YTDnkE1D
π Introducing Qwen3.6-Max-Preview, an early preview of our next flagship model HighlightsοΌ β‘οΈ Improved agentic coding capability over Qwen3.6-Plus π Stronger world knowledge and instruction following π Improved real-world agent and knowledge reliability performance Smarter, https://t.co/0Fr8jgqDbJ
π£ Introducing the Qwen-Robot Suite β Qwen-RobotNav, Qwen-RobotManip, Qwen-RobotWorld, three foundation models, a full stack for embodied intelligence. π§ Qwen-RobotNav β the gateway to mobility. β’ Unifies 5 navigation tasks in one model: instruction following, point-goal, https://t.co/noumjTtTeS
π Introducing FlashQLA: high-performance linear attention kernels built on TileLang. β‘ 2β3Γ forward speedup. 2Γ backward speedup. π» Purpose-built for agentic AI on your personal devices. π‘Key insights: 1. Gate-driven automatic intra-card CP. 2. Hardware-friendly algebraic https://t.co/pA9HCHwFZw
