Cost-Optimal LLM Routing with Limited User Feedback under User Satisfaction Guarantees
For LLM service providers, SLARouter offers a practical solution to balance cost and quality under SLA constraints without requiring extensive tuning or complete feedback.
SLARouter is an online LLM routing algorithm that learns a cost-optimal policy from sparse, one-sided user feedback while providing theoretical guarantees for cost optimality and strict SLA compliance, reducing operating cost by up to 2.2x over baselines.
Inference costs for large language model (LLM) applications are rapidly growing, driven by surging demand and rising infrastructure cost. Users expect high-quality responses, and in commercial settings this is formally codified in Service Level Agreements (SLAs), creating a fundamental tension between cost and quality. Recent progress on cost-aware LLM request routing has shown potential to resolve this tension, but existing approaches rely on complete feedback signals, offline training, extensive per-workload tuning, and most lack SLA guarantees or inference-time adaptivity. We introduce SLARouter, an online routing algorithm that learns a cost-optimal policy from the sparse, one-sided user feedback available in production systems. SLARouter provides theoretical guarantees for both cost optimality and strict SLA compliance. Experiments across a wide range of LLM benchmarks show that SLARouter satisfies SLA constraints without the need for per-benchmark tuning, reducing operating cost by up to 2.2x over existing baselines.