mlsanityData to 7 Oct 2026

LLM Routing as Reasoning: A MaxSAT View

arXiv 2603.13612 ↗First posted 13 Mar 2026cs.AIcited by 0

On the map: LLM reasoning → Test-time compute and efficient thinking

Abstract (arXiv)

Routing a query through an appropriate LLM is challenging, particularly when user preferences are expressed in natural language and model attributes are only partially observable. We propose a constraint-based interpretation of language-conditioned LLM routing, formulating it as a weighted MaxSAT/MaxSMT problem in which natural language feedback induces hard and soft constraints over model attributes. Under this view, routing corresponds to selecting models that approximately maximize satisfaction of feedback-conditioned clauses. Empirical analysis on a 25-model benchmark shows that language feedback produces near-feasible recommendation sets, while no-feedback scenarios reveal systematic priors. Our results suggest that LLM routing can be understood as structured constraint optimization under language-conditioned preferences.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning. Builds on 8 of its 14 references that are in the map. First posted on arXiv 13 Mar 2026.

Builds on

8 of 14 references.

References · 8

  1. 2023FrugalGPT: How to Use Large Language Models While Reducing Cost and Improving Performancecited by 791
  2. 2022Quantifying Memorization Across Neural Language Modelscited by 755
  3. 2021Deduplicating Training Data Makes Language Models Bettercited by 705
  4. 2024RouteLLM: Learning to Route LLMs with Preference Datacited by 633
  5. 2021Documenting Large Webtext Corpora: A Case Study on the Colossal Clean Crawled Corpuscited by 481
  6. 2020Learning under Concept Drift: A Reviewcited by 470
  7. 2024Hybrid LLM: Cost-Efficient and Quality-Aware Query Routingcited by 354
  8. 2024Modeling Multi-Objective Tradeoffs with Monotonic Utility Functionscited by 2

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025Adaptive LLM Routing under Budget Constraints
  2. 2025Learning to Route LLMs from Bandit Feedback: One Policy, Many Trade-offs
  3. 2026Dynamic Model Routing and Cascading for Efficient LLM Inference: A Survey
  4. 2026Learning to Route LLMs from Implicit Cost-Performance Preferences via Meta-Learning
  5. 2025LLM Bandit: Cost-Efficient LLM Generation via Preference-Conditioned Dynamic Routing
  6. 2024Harnessing the Power of Multiple Minds: Lessons Learned from LLM Routing
  7. 2025Constraint-Aware Route Recommendation from Natural Language via Hierarchical LLM Agents
  8. 2026Think When Needed: Model-Aware Reasoning Routing for LLM-based Ranking