Open access
Jul 2026
Contamination-Free LLM Routing on LiveBench Reasoning Tasks: Accuracy-Cost-Latency Tradeoff Learning
Fresh, objectively scored benchmark items can support auditable accuracy-cost-latency routing when features encode verifiable computational structure, and show that fresh, objectively scored benchmark items can support auditable accuracy-cost-latency routing when features encode verifiable computational structure.
Grace Xu
· Journal of Science Innovatio... · 0 citations