UniScale: Adaptive Unified Inference Scaling via Online Joint Optimization of Model Routing and Test-Time Scaling

Bridging model routing and test-time scaling for smarter, more efficient LLM inference deployments.

In real-world deployments of large language models (LLMs), balancing inference quality and computational cost has become a central challenge. Existing approaches tackle this trade-off along two largely independent dimensions: model routing, which switches among models of differen...

ArXiv CS.AI2026-06-01