Tencent Hunyuan Releases Hy Memory 2.0: Fully Open-Sourced Across Model, Framework, and Application Layers, Ranking First in Open-Source Memory Framework Evaluations

Beating AI News Flash: Tencent Hunyuan announced that its agent memory framework Hy Memory has been upgraded to 2.0, and has open-sourced the dedicated model Hy-Memory-30A3B, the memory framework, and the MemHub application. The three-layer capabilities can be used independently or combined into a complete Memory solution. The framework handles the storage, updating, and evolution of three types of memory—chat, work, and code—and ranks first among open-source memory frameworks on 4 Benchmarks; the dedicated model scored 55.49 in evaluations across multiple memory systems including Hy Memory, Mem0, and MemOS, surpassing Opus5's 54.62.

Hy Memory 2.0 upgrades chat memory from scattered facts to complete event threads, and adds work memory and code memory, enabling it to determine the applicable conditions of experience and precisely assist long-horizon tasks. As a local memory management application, MemHub supports CLI and Mac, can connect to 10 AI tools including WorkBuddy, Cursor, and Codex, and enables historical memory import and real-time cross-tool synchronization. This capability has already been deployed in businesses such as Tencent Self-Selected Stocks, iWiki, and Tencent News: for Self-Selected Stocks, multi-hop recall increased from 50% to 90%, and the comprehensive score rose by 23.8%; for iWiki, the accuracy of answering using only memory improved by 48.6% relative to providing all original conversations; for Tencent News, the average score in long-term preference evaluations was 91.68.