Factory has unveiled Droid 1.0, a model-independent coding agent designed to streamline programmatic tool calling as its default operation mode. This innovative harness is aimed at enhancing efficiency in coding tasks across various applications.
In testing on Factory’s Terminal-Bench 4, Droid 1.0 achieved an impressive score of 65.2% when paired with Claude Opus 5.5. The cost per passed task was recorded at $4.73, highlighting the agent’s potential for cost-effective performance in coding environments.
The launch of Droid 1.0 represents a significant advancement in AI-driven coding solutions, as it allows developers to utilize a flexible agent that is not tied to a specific model, thereby broadening its applicability in diverse programming scenarios.
Compiled automatically by the Tech AI Newsdesk from public AI-news sources and summarised in our own words.
