Nace AI has announced the open-sourcing of Drex Max v1, a 27 billion parameter multimodal decision model capable of running locally on Apple Silicon. This model has achieved a notable score of 67.52, showcasing its potential for various applications in decision-making tasks.
In a parallel development, DeepSeek has introduced dsh-superagent, a router-worker coding agent that transforms requests from platforms like GitHub and Feishu into parallel worker sessions. In internal simulations, this prototype demonstrated a significant reduction in task completion time, dropping from 29.5 minutes to 15.7 minutes. Both models represent important advancements in the open-source AI landscape, offering new tools for developers and researchers alike.
Compiled automatically by the Tech AI Newsdesk from public AI-news sources and summarised in our own words.
Frequently asked questions
What is Drex Max v1?
Drex Max v1 is a 27 billion parameter multimodal decision model capable of running locally on Apple Silicon.
What improvements does dsh-superagent offer?
The dsh-superagent transforms requests from platforms like GitHub and Feishu into parallel worker sessions, reducing task completion time from 29.5 minutes to 15.7 minutes.
