TETHER AI

BotPsy: Sub-Billion Parameter Agentic Model Optimized for On-Device Use

Tether AI Research introduces BotPsy, a 0.8B parameter model built on Qwen3.5 designed for function calling and local agent workflows.

2026-11-11

The heavy operational costs and privacy vulnerabilities associated with frontier cloud models are driving a major shift toward edge computing. Addressing this demand, Tether AI Research has released BotPsy, a lightweight 0.8B-parameter model built on top of Qwen3.5-0.8B.

Engineered specifically for function calling and complex agentic workflows, BotPsy proves that routine tasks do not require massive frontier intelligence. Operating efficiently on-device, the model offers developers a self-hosted, highly secure, and extremely cost-effective alternative for localized automation without third-party cloud dependencies.