FrontierCode model benchmark

A Sip of Words/cognition.com/frontiercode/v1.1/14 Oct 2026

Pass vs the selected metric

Pass against the selected metric. One line per model connects its effort levels.

Select

Rows
Effort & ranges
Effort
Avg cost
Pass %
Include models

Results

frontier dominated obsolete

Benchmark results, sortable by column.
Model Effort Pass Pass points behind the leading run Avg cost Out tok Steps