mlsanityData to 7 Oct 2026

When Routing Collapses: On the Degenerate Convergence of LLM Routers

arXiv 2602.03478 ↗First posted 3 Feb 2026cs.AIcited by 10

Not on the map. Why?

Abstract (arXiv)

LLM routing aims to achieve a favorable quality--cost trade-off by dynamically assigning easy queries to smaller models and harder queries to stronger ones. However, across both unimodal and multimodal settings, we uncover a pervasive yet underexplored failure mode in existing routers: as the user's cost budget increases, routers systematically default to the most capable and most expensive model even when cheaper models already suffice. As a result, current routers under-utilize small models, wasting computation and monetary cost and undermining the core promise of routing; we term this phenomenon routing collapse. We attribute routing collapse to an objective--decision mismatch: many routers are trained to predict scalar performance scores, whereas routing decisions ultimately depend on discrete comparisons among candidate models. Consequently, small prediction errors can flip relative orderings and trigger suboptimal selections. To bridge this gap, we propose EquiRouter, a decision-aware router that directly learns model rankings, restoring the role of smaller models and mitigating routing collapse. On RouterBench, EquiRouter reduces cost by about 17% at GPT-4-level performance compared to the strongest prior router. Our code is available at https://github.com/AIGNLAI/EquiRouter.

Key facts. Cited by 10 papers in the map (data through 7 Oct 2026). Builds on 25 of its 30 references that are in the map. First posted on arXiv 3 Feb 2026.

Builds on

25 of 30 references.

References · 25

  1. 2019Sentence-BERT: Sentence Embeddings using Siamese BERT-Networkscited by 9,049
  2. 2017FiLM: Visual Reasoning with a General Conditioning Layercited by 2,286
  3. 2025Qwen3 Embedding: Advancing Text Embedding and Reranking Through Foundation Modelscited by 1,175
  4. 2023HuggingGPT: Solving AI Tasks with ChatGPT and its Friends in Hugging Facecited by 1,173
  5. 2024RouteLLM: Learning to Route LLMs with Preference Datacited by 633
  6. 2024Hybrid LLM: Cost-Efficient and Quality-Aware Query Routingcited by 354
  7. 2024RouterBench: A Benchmark for Multi-LLM Routing Systemcited by 222
  8. 2024GraphRouter: A Graph-based Router for LLM Selectionscited by 154
  9. 2023AutoMix: Automatically Mixing Language Modelscited by 116
  10. 2025BEST-Route: Adaptive LLM Routing with Test-Time Optimal Computecited by 63
  11. 2024TensorOpera Router: A Multi-Model Router for Efficient LLM Inferencecited by 59
  12. 2025RouterEval: A Comprehensive Benchmark for Routing LLMs to Explore Model-level Scaling Up in LLMscited by 58
13 more
  1. 2025MixLLM: Dynamic Routing in Mixed Large Language Modelscited by 56
  2. 2024EmbedLLM: Learning Compact Representations of Large Language Modelscited by 46
  3. 2024Harnessing the Power of Multiple Minds: Lessons Learned from LLM Routingcited by 43
  4. 2024What is the Role of Small Models in the LLM Era: A Surveycited by 41
  5. 2025IRT-Router: Effective and Interpretable Multi-LLM Routing via Item Response Theorycited by 40
  6. 2024Routoo: Learning to Route to Large Language Models Effectivelycited by 34
  7. 2024Eagle: Efficient Training-Free Router for Multi-LLM Inferencecited by 34
  8. 2025Beyond GPT-5: Making LLMs Cheaper and Better via Performance-Efficiency Optimized Routingcited by 32
  9. 2025Causal LLM Routing: End-to-End Regret Minimization from Observational Datacited by 27
  10. 2025ICL-Router: In-Context Learned Model Representations for LLM Routingcited by 26
  11. 2025Capability Instruction Tuning: A New Paradigm for Dynamic LLM Routingcited by 25
  12. 2025Leveraging Uncertainty Estimation for Efficient LLM Routingcited by 21
  13. 2026MMR-Bench: A Comprehensive Benchmark for Multimodal LLM Routingcited by 13

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 10.

10 papers.

  1. 2026Dynamic Model Routing and Cascading for Efficient LLM Inference: A Surveycited by 16
  2. 2026From Sampled Outcomes to Capability Distributions: Rethinking Supervision for LLM Routingcited by 3
  3. 2026SeqRoute: Global Budget-Aware Sequential LLM Routing via Offline Reinforcement Learningcited by 2
  4. 2026RouteJudge: An Open Platform for Reproducible and Preference-Aware LLM Routingcited by 1
  5. 2026RouteGuard: Certifying Routing Gain in LLM Multi-Agent Systems When Complementarity Is Not Enoughcited by 1
  6. 2026Control the Harness, Control the Cost: Routing and Governing AI Coding Agents in the Enterprisecited by 1
  7. 2026The Routing Plateau: Understanding the Accuracy Limits of LLM Routerscited by 0
  8. 2026Pretrain Once, Route Anywhere: Towards a Foundation Model for LLM Routingcited by 0
  9. 2026Routing Should Pay for Itself: Sparse Supervision for Economical LLM Routingcited by 0
  10. 2026False Floors: LLM Safety Routing Evaluations Break Under Distribution Shiftcited by 0