CG(@cgtwts):Not Diamond just made coding agent evals way more realistic. They simulate actual user sessions with long context, cache expiry, and messy interactions. Their router gets Opus xhigh-level quality at 20–60% lower cost. this is a pretty big deal.

2026.09.01 20:16

Not Diamond just made coding agent evals way more realistic. They simulate actual user sessions with long context, cache expiry, and messy interactions. Their router gets Opus xhigh-level quality at 20–60% lower cost. this is a pretty big deal.

Tomas Hernando Kofman@tomas_hk

2026.09.01 16:05

Today we’re releasing our methodology for evaluating model routing with interactive benchmarks, which represent agent cost accumulation better than static benchmarks do. Across leading benchmarks, we achieve Pareto-dominance, exceeding Opus xhigh quality at 20–80% lower cost.