coding-model-comparison

Expand benchmark rubric with interaction completeness

Refresh benchmark results and add Kimi 2.6 comparison

Add configurable parallel evaluation with live progress updates

Stabilize evaluation scoring and refine rubric breakdown

Add three model pages and update index & results