EVAL
Public benchmarks
Weekly evaluations of SCR, Claude Sonnet 4, and GPT-4 on a curated private-markets benchmark set. Dataset private; results public.
Frozen at —
MethodologyNo benchmark results yet. The first weekly run populates this table.
EVAL
Weekly evaluations of SCR, Claude Sonnet 4, and GPT-4 on a curated private-markets benchmark set. Dataset private; results public.