AI coding benchmark MirrorCode published its full results June 26, showing Claude Opus 4.7 autonomously rebuilt a 60,000-line interpreter and scored 56% overall — completing tasks that take human ...
dIBTCI, Global Health Technical Assistance and Mission Support Project (GH-TAMS), Washington, DC, United States eUnited States Agency for International Development, Bureau of Global Health, Office of ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results