The authors re-scored 600 publicly released coding-agent trajectories from two agent scaffolds and two open-weight models. Median tool observations were small, 118 to 286 tokens, and the large build logs often used to justify output…
In the News: October 16, 2026 (Midday)
A study of 600 coding-agent runs finds turn count, not tool-output size, drives cost; a new benchmark scores 19 models on building harnesses.
Midday edition
Machine-readable
Download Markdown
Story
A study of 600 coding-agent runs finds turn count explains 84% of cost variance
Read story →
Story
Doc2Harness scores 19 models on building a coding-agent harness; the best passes 53%
Each model is given a development document consolidated from the user documentation of twelve frontier harnesses, plus a skeleton that declares interfaces only, and has to build a complete harness. The frozen harness is then run end to…