Utkarsh-X/Supreme
A lean engineering constitution for coding agents. Evaluated across all 89 Terminal-Bench 2.1 tasks (267 isolated container runs) with full research paper and telemetry.
正在读取项目信息…
A lean engineering constitution for coding agents. Evaluated across all 89 Terminal-Bench 2.1 tasks (267 isolated container runs) with full research paper and telemetry.