Factory Router, from Factory, trades a small amount of benchmarkA standard public test set for comparing AI models — the shared scoreboards behind every "model X beats model Y" claim.Full definition → performance for lower session cost in its Droid coding AI agentAn AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.Full definition →. Against Anthropic's Opus 4.7 baseline, Factory reports 99% of the baseline pass rate at 20% lower cost on Terminal-Bench 2, and 96% at 25% lower cost on Legacy-Bench. These are relative pass rates, not claims that nearly every task succeeded. The router report measures the cost of complete sessions.
The router assigns less expensive models where they can handle the work and can escalate struggling sessions. More aggressive routing reduced both cost and success rates further. The harness explanation supplies architectural context windowThe maximum amount of text a model can consider at once — its working memory for the current conversation or task.Full definition → for these choices.
Terms in this piece · Glossary
benchmark — A standard public test set for comparing AI models — the shared scoreboards behind every "model X beats model Y" claim.
AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
agent harness — The scaffolding around a model that turns it into a working agent — the loop, the tools it can call, and the rules for when to stop.
context window — The maximum amount of text a model can consider at once — its working memory for the current conversation or task.
