Tether releases BotPsy, an 800-million-parameter on-device AI agent model.
Original headlineTether 发布 8 亿参数端侧 AI 智能体模型 BotPsy
Not investment advice. Muhabbit scores how much a story may matter, not which way to trade.
Tether AI Research released QVAC BotPsy, an open-weight 800-million-parameter model built on Alibaba's Qwen3.5 architecture for on-device tool calling and agentic tasks. Tether said BotPsy scored roughly double its base model on the BFCL v4, ACEBench, τ-bench and ToolSandbox benchmarks, and outperformed the 2-billion-parameter Qwen3.5-2B-Think. The model converts user requests into local tool calls, and developers can adapt and deploy its weights in their own systems. The release signals Tether's push beyond stablecoins into open, edge-deployable AI, a segment where small models compete on efficiency rather than scale. What to watch next is developer adoption of the weights and whether BotPsy's benchmark gains hold up in real-world agent deployments.
Muhabbit's AI wrote this summary in its own words from TechFlow's report and translated it from Chinese. It can get details wrong, so check the source.
Named in this story
Tether, QVAC BotPsy, Tether AI Research, Qwen3.5, Alibaba (BABA)
A linked name opens that asset's sentiment page.