Tether AI's BotPsy is a 0.8B parameter model built on Qwen3.5 designed for local function calling. It doubles the agentic performance of its base model on benchmarks like BFCL v4 and beats the larger Qwen3.5-2B on most tool-use tasks. The focus is on handling multi-turn workflows and knowing when not to call tools.
It is worth testing if you need a lightweight, private assistant that can reliably execute simple commands. The catch is that it is explicitly limited to low-to-medium difficulty tasks and makes mistakes. Do not deploy it for high-stakes operations or complex reasoning.
more: https://huggingface.co/blog/qvac/botpsy
my other side: ivo.blogspot.com
Leave a Reply