Operating manual

Tag: terminal-bench

A focused reading path for terminal-bench: related field notes, evidence trails, and operating questions from the archive.

agent-ops/
24 Sep, 2026

compaction-flips-the-model-tier-tradeoff.md

Compaction Flips the Model-Tier Tradeoff

A v1 preprint on cost-efficient compaction for long-horizon coding agents reports >10pp Terminal-Bench uplift for less than the cost of two full-context runs. Pricing compaction at the rollout boundary, not end-to-end, lets a cheaper model match a flagship.
open artifact →