Meet Factory Droid
Factory's move signals where the coding-agent market is heading: the harness, not the model, becomes the product, with frontier models swappable underneath depending on performance and price. Making programmatic tool calling the default fits the broader shift toward script-mode orchestration over chatty JSON tool loops.… more
The 65.2% Terminal-Bench 4 result is respectable, but you should treat it as self-reported — these are Factory's own runs, not independent verification, which is exactly why the $4.73-per-passed-task figure deserves scrutiny. Still, the cost-per-passed-task framing is the right metric to popularize, since raw scores hide how expensive a benchmark pass really is. Expect competitors like Devin and Cursor to counter with their own harness-level benchmarks soon.
Meet Factory Droid 1.0: A Model-Independent Coding Agent Harness Built on Script Mode
Factory released Droid 1.0, a model-independent coding agent harness that makes programmatic tool calling the default. On Factory’s own Terminal-Bench 4 runs, Droid scores 65.2% with Claude Opus 5.5 at $4.73 per passed task. The post Meet Factory Droid 1.0: A Mod…
Source:
Marktechpost
Discussion (0)