[{"id":"hard-h2h-pass-matrix","title":"Pass matrix on hard tasks: configuration × task (strict passes)","columns":{"$k":["key","label","unit"],"$r":[["config","Configuration","text"],["c0","Fix an interval-merge function (off-by-one and edge cases)","text"],["c1","Fix a time-zone day-length function (DST)","text"],["c2","Write a CSV parser (quoted newlines, strict errors)","text"],["c3","Predict JavaScript event-loop output order","text"],["c4","Solve a multi-constraint room schedule","text"],["c5","Write a strict SemVer 2.0.0 regex","text"],["c6","Refactor to remove duplication, keep 20 tests green","text"],["c7","Write a SQLite reporting query (fan-out, ties, boundaries)","text"],["strict","Strict total","text"],["formatMisses","Format misses","count"],["wrong","Wrong answers","count"],["medianTotal","Median total (s)","seconds"]]},"rows":{"$k":["config","strict","formatMisses","wrong","medianTotal","c0","c1","c2","c3","c4","c5","c6","c7"],"$r":[["Claude Sonnet 5.5 · Claude Code","24/24",0,0,7.75,"3/3","3/3","3/3","3/3","3/3","3/3","3/3","3/3"],["Claude Opus 5.5 · Claude Code","24/24",0,0,9.18,"3/3","3/3","3/3","3/3","3/3","3/3","3/3","3/3"],["Claude Opus 5.5 (high) · Claude Code","24/24",0,0,11.03,"3/3","3/3","3/3","3/3","3/3","3/3","3/3","3/3"],["GPT-6.1 Sol (medium) · Codex CLI","16/16",0,0,13.11,"2/2","2/2","2/2","2/2","2/2","2/2","2/2","2/2"],["Claude Fable 5.1 · Claude Code","24/24",0,0,16.13,"3/3","3/3","3/3","3/3","3/3","3/3","3/3","3/3"],["GPT-6.1 Sol (high) · Codex CLI","16/16",0,0,18.12,"2/2","2/2","2/2","2/2","2/2","2/2","2/2","2/2"],["Claude Haiku 4.5 · Claude Code","11/24",5,8,39.01,"3/3","1/3","2/3 (+1 format miss)","0/3","0/3 (+2 format misses)","3/3","2/3","0/3 (+2 format misses)"]]}},{"id":"hard-h2h-controls","title":"Validator controls run before the first model call","columns":{"$k":["key","label","unit"],"$r":[["task","Task","text"],["reference","Reference answer","text"],["checks","Checks","count"],["wrong","Plausible wrong answers rejected","text"],["wrapped","Wrapped reference flagged as format miss","text"]]},"rows":{"$k":["task","reference","checks","wrong","wrapped"],"$r":[["Fix an interval-merge function (off-by-one and edge cases)","passes",18,"5/5","yes"],["Fix a time-zone day-length function (DST)","passes",36,"3/3","yes"],["Write a CSV parser (quoted newlines, strict errors)","passes",22,"3/3","yes"],["Predict JavaScript event-loop output order","passes",1,"3/3","yes"],["Solve a multi-constraint room schedule","passes",14,"3/3","yes"],["Write a strict SemVer 2.0.0 regex","passes",30,"3/3","yes"],["Refactor to remove duplication, keep 20 tests green","passes",25,"2/2","yes"],["Write a SQLite reporting query (fan-out, ties, boundaries)","passes",8,"4/4","yes"]]}}]