Aftermarket Harnesses

Tomasz Tunguz ·

Benchmark results show coding-agent harnesses can dominate model performance and token costs through context management and caching.

Categories: Research

Excerpt

The harness now moves the coding benchmark more than the model does. Endor Labs' Agent Security League found GPT-5.5 scored 61.5% functional correctness in Codex & 87.2% in Cursor, & Claude Opus 4.7 scored 87.2% in Claude Code & 91.1% in Cursor. Input tokens are 86-98% of OpenRouter volume, so the harness controls most of the bill through cache discipline. First-party co-design buys real cache hit rates, but a third-party harness can match them.