LLM Routing

LLM: Large Language Model

Follow this topic Ask about this topic

Momentum

30 papers in the last four weeks, up 88% on the four weeks before. 0.3% of all new papers.

Jul 13Week of Sep 28

Latest papers 224

Oct 7, 2026cs.CR

Sensitive-Topic Leakage Through LLM Routing Metadata: Measurement and Mitigation

LLM routers pick a cheap or expensive model per request by its content, and many gateways and some cloud platforms can log that choice with content logging off. We measure this privacy channel beyond token counts, accounting for noisy labels and repeated prompts. We run pre-registered studies on 1.7 million real requests (WildChat-1M, LMSYS-Chat-1M) with two cost/quality routers and a domain router, survey eleven systems' logging, and test post-processing defenses. At matched length, the shift's direction depends on category and router. For RouteLLM at the 50% operating point, harassment and self-harm requests reach the strong model 19 points less often than comparable ones on prompts unseen in exploration, medical requests (exploratory: LLM labels failed their gate) 31 points less often on distinct prompts (both post hoc), and sexual requests 10 points more often (secondary); the other router's four are negative. Twenty RouteLLM decisions separate frequent medical askers with AUC 0.71, exploratory and below the pre-registered primary endpoint's 0.75 (domain router: 0.92, an upper estimate). Per-category length-matched parity with accurate labels removes the gap on real traffic, costing at most 0.2 accuracy points on RouterBench (post hoc), where routers' gaps on sensitive subjects (13-42 points, pre-registered) exceed those of an oracle routing by realized accuracy gain (1-11, post hoc). Per-conversation stickiness, per-user budget bands, and pooled parity fail, the last as categories' shifts differ in size or sign. A post hoc exact per-user rate hides only even-prefix strong counts and forfeits most self-assessed routing value; it preserves odd-position decisions, from which a post hoc log attack reaches AUC 0.73 after 20 RouteLLM requests (exploratory).

Teng-Ruei Chen

Krixvon, Taipei 100, Taiwan

Privacy Leakage in Language Models LLM Routing LLM Security Privacy-Preserving Language Models LLM Auditing

Oct 7, 2026cs.LG

Evaluating Trajectory Features for Routing Final-Layer Attention

Attention routing requires a signal that predicts the value of attention on the current prefix. We evaluate whether hidden-state extrapolation error, curvature and error change improve this prediction beyond uncertainty, one-step displacement, position and state projections. Paired executions of the final attention layer supply signed next-token loss differences in frozen SmolLM3-3B-Base and Qwen3.5-4B-Base checkpoints. Utility-supervised routers are tested on 100 held-out PG-19 books at an identical causal 20 percent invocation quota. None of six prespecified comparisons shows a positive gain after familywise correction. In Qwen3.5, a parameter-matched fixed-projection control lowers NLL by 0.00356 nats/token relative to the trajectory router (95 percent interval 0.00218 to 0.00487). Secondary results depend on the operation removed, feature location and scoring horizon; frozen thresholds also drift substantially at longer horizons. Actual selected-query execution yields small long-sequence latency reductions with increased NLL, while learned routers remain slower during cached continuation. The study identifies limits on the incremental value of these trajectory summaries and separates allocation quality from measured inference benefit.

Yupeng Yao

LLM Routing LLM Inference Acceleration Long-Context Language Model Inference Sparse Attention

Oct 6, 2026cs.AI

Not Every Call Needs a Frontier Model: Per-Call-Site Evaluation of Small Language Models in a Deployed Agentic Home-Automation System

An agentic system issues several structurally different kinds of LLM calls. It routes intent, classifies actions, grounds language in a device registry, plans multi-agent pipelines and writes the Python code those pipelines run. The difficulty of these call sites varies by an order of magnitude, yet in practice a single model, chosen for the hardest site, serves all of them. In this work, we evaluate 9 models from 0.8B to a frontier hosted model across the five call sites of a deployed open-source home-automation framework (Wactorz), using its unmodified production prompts and two real Home Assistant installations (280 cases, 2520 scored calls). We find that capability is not ordered the same way at every site, and that larger models are not uniformly better: one 4B model is worse than its 2B sibling at grounded actuation. Paired testing shows the best local model to be statistically indistinguishable from both hosted models at four of five sites. Only code generation separates them, against a small hosted model (p = 0.039) as well as a frontier one (p = 0.002). Aggregate accuracy also hides a safety failure specific to actuation, where small models resolve the accuracy/refusal trade-off in degenerate ways: one model (Gemma4 E2B) actuates on 87.2% of requests for devices the site does not own, while another refuses every request it receives. Routing each site to its best local model reaches 91.8% against 95.4% at no per-call cost. In a live deployment judged by a user, hosting only the two generative sites matches hosting everything (39/43 against 39/43) for 28% of the spend, and the actuation gap the benchmark predicted appears as exactly one case in twenty-six. Benchmark, harness and all records are released at https://github.com/waldiez/slm-callsite-eval.

Panagiotis Kasnesis, Christos Chatzigeorgiou, Lazaros Toumanidis +1

University of West Attica, Athens, Greece · Waldiez PC · ThinGenious PC

Small Language Models LLM Agent Evaluation LLM Routing Language Model Safety Evaluation

Oct 6, 2026cs.AI

Do I Need the Cloud? Uncertainty-Aware Step-Level Handoff for Small Language Model Agents

Small language models (SLMs) are attractive as local agent controllers because they reduce remote inference, latency, and deployment footprint, yet structured tool errors can cause an agent step to fail. Existing routers typically select a model once per query. However, agents expose sequential decision points whose difficulty dynamically changes based on intermediate observations. We propose STEPGATE, an uncertainty-aware handoff framework that scores each local SLM action and selectively escalates challenging steps to a stronger model. On a 52-task held-out single-step BFCL-derived test split, the Qwen2.5-1.5B/7B pair attains 82.7% task success with 30.8% escalation, versus 67.3% local-only and 75.4% random escalation (which uses 33.8% escalation). In a separate multi-turn evaluation, STEPGATE achieves 69.0% trajectory success and 84.0% action success using only 30.0% cloud actions, compared with 48.0%/70.5% local-only, 60.0%/78.2% random escalation, and 57.0%/77.1% query-level routing (strong-only achieves 82.0% trajectory success at 100% cloud actions). These results suggest that step-level escalation recovers a large share of the performance gap to the stronger Qwen2.5-7B backend at a matched cloud-action rate while transmitting fewer tokens remotely. However, our evaluation is limited to one model family, a single stronger backend, and scripted tasks. Furthermore, the test sets are small, multi-turn comparisons rely on paired intervals and statistical tests, and our risk tiers serve as research annotations rather than formal safety guarantees.

Abolfazl Younesi

Department of Computer Engineering Sharif University of Technology Tehran, Iran

Small Language Models LLM Agents LLM Routing LLM Agent Reliability

Oct 5, 2026cs.LG

Stepped MoE: Segment-Level Routing with Configurable Inference Complexity

Training large language models (LLMs) is resource-intensive, and adapting them for diverse deployment scenarios with varying computational constraints remains challenging. While elastic architectures enable flexible model deployment and sparsely activated models allow input-adaptive computation, existing approaches treat these dimensions independently. Moreover, models catered towards on-device edge inference need to conform to the memory and compute limitations of the serving devices. In this paper, we introduce a unified framework that combines elastic structures with sparsely gated architectures to create models that adapt simultaneously to both deployment constraints and task requirements. Our approach employs a model backbone that conditions on both the context and target efficiency specifications, enabling fine-grained control over the accuracy-efficiency trade-off at inference time. The model learns to activate task-relevant parameters within elastically-nested sub-networks, allowing a single model to span multiple capacity points while maintaining input-adaptive routing. Through experiments we demonstrate that we can create a model that allows the flexibility to use 1,2,3,4 billion parameters while being more accurate than their dense counter-parts (2-5% on knowledge-intensive benchmarks) and at par with their static versions while delivering similar latency metrics as dense models. Overall, we save on device disk space by sharing the model parameters, allow flexibility of serving based on DRAM and compute available while delivering more accurate results.

Arnav Kundu, Zhaoyang Xu, Bairu Hou +3

Apple

On-Device Language Model Inference Dynamic Neural Networks Mixture of Experts LLM Routing Efficient Language Model Inference

Oct 5, 2026cs.CL

Cross-lingual Calibration of Pre-Generation Success Probes for Multilingual LLM Routing

Pre-generation success probes estimate response correctness from a language model's hidden activations before decoding, enabling cost-aware routing. While prior work has demonstrated their utility primarily on English inputs, we study their reliability across languages along three dimensions: (1) whether they preserve the ranking of likely successes and failures (DISCRIMINATION); (2) whether they retain probabilities that match observed success frequencies (CALIBRATION); and (3) whether they produce scores comparable enough across candidate models for cost-aware multilingual routing (UTILITY). Using 3,000 MATH problems in 10 languages and 8 open-weight model configurations, we compare cross-lingual transfer from English-trained probes and equal-budget pooled multilingual probes. English-trained probes retain useful cross-lingual discrimination but become less well calibrated after transfer. Pooled multilingual supervision improves both properties and yields more reliable estimates of success. In routing experiments, the pooled router achieves a 0.7% higher test success rate while reducing modeled cost by 13.0% relative to always selecting the model with the highest average success. These results show that multilingual routing requires success estimates that remain well calibrated and comparable across languages and models.

Andrea Paganelli, Stefano Civelli, Pietro Bernardelle +1

The University of Queensland · Polytechnic University of Milan

Confidence Estimation in Language Models Language Model Calibration LLM Routing Language Model Probing Cross-Lingual Transfer

Oct 5, 2026cs.CL

Breaking the Tie: A Cluster-Aware Routing Framework for Large Language Models

With the rapid development of artificial intelligence, the emergence of various Large Language Models (LLMs) has created a rich model ecosystem. However, this also brings a key challenge: how to select the optimal model for a specific user query. LLM routing addresses this need by dynamically assigning queries to the most suitable expert in the pool of candidate models. However, existing routing frameworks often simplify this process to a standard classification task; thus, a critical vulnerability is exposed when multiple candidate models correctly answer the same query. We formalize this capability overlap as routing noise, which misleads the router with arbitrarily correct candidate models, ultimately leading to routing collapse (a severe decline in generalization ability on unseen tasks). To address this problem, we propose a novel Cluster-Aware Soft-Labeling Routing (CASLR) framework. CASLR shifts the evaluation paradigm from the success of a single query to macro-domain consensus by replacing traditional one-hot vectors with a masked softmax mechanism. Specifically, for experts who answer incorrectly, we penalize their target probability to zero; for the remaining candidates, we directly compute continuous fine-grained soft labels based on their global clustering utility scores. We then use these refined soft labels to supervise a lightweight router. Specifically, the framework not only demonstrates superior accuracy on multiple benchmarks, but also outperforms Llama-3.3-70B-Instruct by 7.80% in overall average performance. Furthermore, the extremely low routing inference latency of only 1.13s further confirms that CASLR can achieve efficient system scheduling with almost zero additional overhead, while ensuring high response quality.

Yao Lu, Zhaiyuan Ji, Yaxin Gao +6

Institute of Cyberspace Security, Zhejiang University of Technology · Binjiang Institute of Artificial Intelligence, Zhejiang University of Technology · Hong Kong University of Science and Technology (Guangzhou)

Oct 1, 2026cs.AI

OR for AI That Does OR: Routing LLMs up the Escalator inside the OSCAR Framework

Large language models can translate business descriptions into optimization models, but executable code may misrepresent constraints or objectives. A solver can then return an optimal solution to the wrong problem. Even when the solution satisfies the intended operating rules, a better plan may exist. For organizations that repeatedly use optimization modeling, an LLM-based framework should produce accurate formulations at low cost and, ideally, run locally. We study how to verify improvements and allocate attempts across LLMs that differ in price and capability. We develop OSCAR (Optimization modeling by Simulator, Coder, And Reviewer), which uses an offline Simulator certified against labeled decision examples to compare candidates and continues searching beyond feasibility. We model the search for the next certified improvement as sequential decisions under unobserved difficulty: which LLMs to call and when to stop. In a simplified known-prior setting, we give conditions under which cost-ordered escalation is optimal. For general menus, we derive a prior-free competitive guarantee. On five benchmark problems, OSCAR achieves 95% to 100% accuracy at the reported settings using two small open-weight LLMs, each deployable locally on a single GPU. Their single-attempt accuracies average 29% and 48%. In five runs per problem, Codex and Claude Code incur average token costs 3.1 and 5.8 times OSCAR's, respectively. OSCAR supports open-weight models locally or in the cloud, depending on budget and confidentiality requirements. Firms should maintain labeled decision examples of feasible and infeasible decisions to clarify plain-language operating rules. OSCAR follows these labels when an LLM's interpretation conflicts with them. As LLM capabilities and prices change, OSCAR's simple operating rules and adjustable settings help firms adapt their model choices and benefit from these advances.

Jinzhi Bu, Haixin Tang, Huanan Zhang

Department of Logistics and Maritime Studies, The Hong Kong Polytechnic University, Hong Kong · Leeds School of Business, University of Colorado Boulder, Boulder, CO, 80309, USA

Large Language Model-Guided Optimization Optimal Stopping LLM Routing Natural Language to Optimization Modeling

Sep 30, 2026cs.SE

Adaptive-GEPA: Make Your Harness Fit Heterogeneous Requests

Reflective optimizers such as GEPA improve language model prompts from execution traces and evaluator feedback; full-program extensions can also rewrite tools and control flow. In practice, a user hands the same endpoint heterogeneous requests whose effective solutions require different tools, reasoning modes, and control flow. Optimizing one shared program leaves this division of work implicit in source-code search, while optimizing a separate program per request family fixes it beforehand. We introduce Adaptive-GEPA, which learns both how to divide requests and how to solve them. It evolves a router and a library of specialist programs under one search budget. The router's instructions, each specialist's description, and its program code are plain, human-readable text, edited from feedback. To combine branches, it aligns specialists by the requests they handle and inherits descriptions together with programs. On a fixed mixture of four task families, the reported Qwen3-8B run evolves four experts without supplying family labels to the router or reflection model; its routing matches the task partition on all 651 test requests. Its family-mean test score (x100) rises from 52.6 to 70.6, compared with 62.5 for GEPA's full-program adapter and 54.0 for GRPO at a nominal budget of 18,000 scored calls. These counts do not equate total compute. Figure 1 summarizes the learning curves, final test scores, and routing agreement.

Tianyu Chen, Yasi Zhang, Ruiyi Wang +3

University of Texas at Austin · University of California, Los Angeles · University of California, San Diego +3

Large Language Model-Guided Optimization LLM-Based Program Synthesis Agent Harness Optimization LLM Routing

Sep 29, 2026cs.LG

FlexRouter: Learning Complementary Model Sets for Flexible LLM Routing

Existing Large Language Model (LLM) routing methods score LLMs independently to select top-

k

models. However, this ignores model correlations and enforces a rigid computational budget. Consequently, routers often select redundant models that share failure modes, limiting the overall probability of success. To address this, we propose FlexRouter, a routing framework that explicitly models model complementarity. FlexRouter optimizes for \textit{answer coverage}, maximizing the probability that at least one selected model yields a correct response. This objective aligns with practical inference pipelines where multiple candidate outputs are generated and a downstream verifier or user selects the final one. We formulate routing as a coverage-oriented subset selection problem and model the routing policy using Determinantal Point Processes (DPPs), which naturally capture both model competence and redundancy. To directly optimize coverage without requiring a ground-truth target subset, we introduce a training objective based on marginalizing over failure sets. During inference, we employ a greedy strategy based on marginal log-determinant gains, enabling the router to adaptively determine subset sizes without a predefined budget. Extensive experiments on the large-scale RouterEval benchmark demonstrate that our proposed FlexRouter achieves higher coverage with lower redundancy across both in-domain and out-of-domain tasks than strong baselines while maintaining flexible inference cost.

Wang Wei, Harry Yang, Tiankai Yang +6

Virginia Tech · Independent Researcher · University of Southern California +2

Adaptive Model Routing LLM Routing

Sep 29, 2026cs.AI

You Cannot Pick a Provider From the Price List: Market-Aware Routing for Open-Weight LLM Inference

Existing LLM routers choose among models using static per-model costs. We show that open-weight inference markets introduce a second, largely ignored decision axis: after choosing a model, a client must still choose which provider serves it. Measuring live endpoints across [nummodels] open models, competing providers, multiple task types, and three measurement waves, we find that provider choice cannot be inferred from the price list. The same model can vary sharply in quality, latency, availability, and price across providers; higher-priced providers are consistently faster, but price does not reliably predict quality or availability; and provider feasibility is task-selective, with one deployment nearly normal on knowledge tasks but catastrophically degraded on multi-step reasoning. We formulate same-model provider selection as a price-taker market-aware routing problem. A simple measured-map policy routes to the cheapest provider that is both quality-equivalent and healthy, yielding matched-quality savings while avoiding degraded endpoints. Because the map drifts, we introduce FACET, an online provider router that certifies per-(provider x task) feasibility facets and fails safe to an anchor before serving uncertified arms. Across relaxed deployment assumptions, FACET tolerates imperfect task assignment and sparse feedback, while systematic evaluator bias exposes a quality-signal trust boundary that can be mitigated with ground-truth probes or audits. Live provider runs further confirm that certification can move real traffic from a premium anchor to a substantially cheaper certified endpoint. Our results suggest that market-aware LLM routing must measure not only which model to use, but also who serves it.

Liang He, Jingbo Wen, Yixiong Chen +4

University of Sydney · Johns Hopkins University · Stanford University +1

LLM Serving Cost-Aware Inference LLM Routing

Sep 29, 2026cs.AI

Cross-Entropy Guided Routing in Mixture-of-Experts Large Language Models

Sparse mixture-of-experts (MoE) large language models scale model capacity by routing each token to a small subset of experts. Their routers are regularized with load balancing terms and learn affinity scores through the language-model objective. However, these objectives do not provide direct alignment between routing affinities and token-level error. We introduce token-error supervision for sparse routing in two forms. The first form predicts an error score per expert. The affinity-weighted aggregate of these scores is aligned to the next-token cross-entropy loss, while the individual scores attenuate affinity before top-

K

selection. The second directly aligns the router's affinities to the model's objective without requiring an additional head or inference-time modification. Both formulations use the Itakura--Saito divergence or an exponential negative log-likelihood for aligning affinities and token errors. Across two sparse MoE backbones and four multiple-choice question-answering benchmarks, we evaluate both supervision mechanisms. On Granite, our method improves accuracy by approximately 2.3 percentage points on average over a parameter-matched routing baseline. With stronger supervision, the gain on ARC-Challenge reaches 2.94 points. Both mechanisms preserve the native sparse execution budget and aggregation policy. Our code is available in the supplementary materials.

Yury Nahshan, Nati Daniel, Jacob Goldberger +1

Bar-Ilan University, Ramat-Gan, Israel · NVIDIA, Israel

Mixture-of-Experts Language Models Parameter-Free Mixture-of-Experts Routing Sparse Mixture-of-Experts Expert Routing LLM Routing

Sep 29, 2026cs.AI

Routing Should Pay for Itself: Sparse Supervision for Economical LLM Routing

Large language model (LLM) routing reduces serving cost by assigning each query to an appropriate model while preserving response quality. Learning such a router, however, often requires executing multiple candidate models on historical queries to collect query--model quality feedback, creating a nontrivial supervision cost before deployment. Existing work largely focuses on serving-time efficiency, overlooking whether the resulting savings are sufficient to recover this upfront expenditure. We further observe that routing quality often saturates well before all query--model feedback is collected, suggesting that dense supervision can be economically over-provisioned. We propose SaveRouter, a sparse-supervision routing framework that selectively acquires informative model feedback and shares capability information across related queries, while retaining query-level refinement for fine-grained routing. We evaluate routing by jointly accounting for supervision expenditure and subsequent serving-time savings. Across four routing benchmarks, the main setting uses only about 33--41% of available training feedback while maintaining competitive or better routing quality, and reduces the break-even deployment volume by approximately 1.9--9.5 times compared with the fastest conventional router. Further analysis shows that acquiring more supervision is not always economically preferable: the supervision level that minimizes serving cost can differ from the one that achieves the earliest payback. Our code is publicly available at https://github.com/LAMDA-Model-Reuse/SaveRouter.

Guannan Lai, Gelin Bian, Hao-Xuan Ma +5

School of Artificial Intelligence, Nanjing University · National Key Laboratory for Novel Software Technology, Nanjing University · SinapisAI +1

LLM Serving Cost-Aware Inference Active Learning LLM Routing

Sep 29, 2026cs.AI

Pretrain Once, Route Anywhere: Towards a Foundation Model for LLM Routing

Large language model (LLM) routing aims to assign each query to the most suitable model from a heterogeneous candidate pool, improving the quality--efficiency trade-off of LLM inference. Existing routers are typically learned through local fitting: a router is optimized for a particular query workload and candidate pool, and often requires additional supervision or retraining as the routing environment changes. We ask whether LLM routing can instead be approached from a foundation-model perspective, learning a reusable routing capability that generalizes across tasks, candidate models, and deployment conditions. To this end, we introduce RouteFM, which learns to characterize anonymous candidate models from behavioral context and infer their target-specific capabilities, rather than binding routing decisions to fixed model identities or a single environment. Through episodic pretraining across heterogeneous routing environments, this capability can be reused by a frozen router and adapted to new environments through context alone. Experiments demonstrate transfer across changes in domains, modalities, candidate pools, and context budgets, with the largest gains when behavioral evidence is limited. On MMR-Bench, which is excluded from pretraining, RouteFM outperforms the strongest baseline by 2.23 quality points with only eight observations per candidate. These results support moving LLM routing from repeated local fitting toward a pretrain once, route anywhere paradigm. Our code is publicly available at https://github.com/LAMDA-Model-Reuse/RouteFM.

Guannan Lai, Han-Jia Ye

School of Artificial Intelligence, Nanjing University · National Key Laboratory for Novel Software Technology, Nanjing University

Sep 28, 2026cs.SE

Evaluating Name-Only Directory Routing for One-Shot Code Search

Finding the right files is an early challenge for coding agents. We test whether a language model can follow directory and file names to find annotated code files missed by fixed lexical queries. Across 82 audited issues from 11 repositories at pinned pre-fix commits, name-only directory routing recovered 0.465 of gold files within eight candidates, compared with 0.352 for FTS5 and 0.245 for a fixed full-issue rg query. The paired gain over FTS5 was 0.113 (95% repository-cluster bootstrap interval, 0.053 to 0.168). Under a shared 16K-token context budget, routing delivered 0.443 of annotated lines versus 0.246 for FTS5 on 55 cases with fully aligned annotations. At the same eight-file limit, combining routing with FTS5 reached 0.491 file recall, but its gain over routing alone was uncertain. An exploratory flat path control reached 0.572 recall while using 24.6 model calls per issue, compared with 8.9 for routing. Routing averaged 8.9 seconds per issue; FTS5 took 7 milliseconds per query after a 0.9-second build. On this cohort, directory routing added relevant file candidates to one-shot lexical search, but the study cannot attribute the gain to hierarchy or show that it improves issue resolution.

Manoj Bajaj

Information Retrieval LLM Routing IR Evaluation

Sep 28, 2026cs.AI

SeLMRoute: Probabilistic Semantic Evidence for Large Language Model Routing

Large language model (LLM) routing aims to select the most suitable model for each incoming query. Most existing routers learn this decision directly from query embeddings, model representations, preference data, or clusters of similar examples. Such approaches can be effective, yet the representation used for routing rarely states what a query actually requires. We introduce SeLMRoute, a routing framework that separates the extraction of candidate-independent semantic evidence from the learning of candidate performance and the application of deployment objectives. A decision model first evaluates a set of interpretable questions about the query, such as its reasoning requirements and use of external knowledge, with each judgment retained as a probability distribution. The resulting probabilistic semantic state is used by a lightweight supervised router to estimate candidate model performance. Routing objectives are applied after performance estimation, which allows the same semantic state to support performance-oriented and cost-aware decisions. On the LLMRouterBench (15 datasets, 20 candidate models, 11,481 queries), SeLMRoute achieves an average accuracy of

72.08\% \pm 0.45

, while grouped five-fold out-of-fold evaluation reaches

72.64\%

, compared with

69.23\%

for the strongest fixed candidate. The representation achieves the highest mean performance among the evaluated semantic, dense, lexical, and domain-level representations. In a separate 13-model performance-cost setting, SeLMRoute improves performance in all five grouped splits, with a mean PerfGain of

2.66\%

. Our code is available at https://github.com/Indigma-Innovations/SeLMRoute.

Vasilis Perifanis, Nikolaos Pavlidis, Symeon Symeonidis

Indigma Innovations · Democritus University of Thrace · Athena Research Center

Sep 28, 2026cs.AI

RSI-Router: Evolving Subtask-Level LLM Routing and Skills for Cost-Efficient Agents

Practical deployment of large language model (LLM) agents requires strong task performance at affordable inference cost. For long-horizon agentic tasks, this performance-cost trade-off can be improved through within-task large-small model collaboration, as smaller models can handle some stages even when they cannot solve the full task. In this paper, we introduce RSI-router, a routing framework that constructs subtask-level model assignments and model-specific skills through recursive self-improvement over accumulated experience. Each iteration consists of four stages: Subtask Mining derives subtask definitions and identification rules from training trajectories; Routing Strategy Evolution proposes and evaluates diverse model assignments; Model-Specific Skill Evolution compares routed and large-model-only trajectories to diagnose failures and develop reusable execution skills; and Pareto-Optimal Router Selection updates the Pareto population using historical and newly generated routers while retaining dominated routers as experience for subsequent evolution. Routing between DeepSeek-V4.1-Flash and Qwen3.5-9B, RSI-router consistently surpasses the DeepSeek-only baseline at roughly half the inference cost (48.3%) across five agentic benchmarks. In particular, on ALFWorld, ScienceWorld, and WebShop, it cuts inference cost by 74.7-82.2% while simultaneously improving performance; on Terminal-Bench 2.0, it achieves a 16.7% relative performance gain at 18.0% lower cost. Moreover, RSI-router establishes a stronger performance--cost Pareto frontier than 9 routing methods.

Hao Li, Hangfan Zhang, Zhiyao Cui +5

Shanghai Artificial Intelligence Laboratory · Northwestern Polytechnical University

Cost-Aware Inference LLM Agent Workflow Optimization LLM Agent Skill Learning LLM Routing Long-Horizon LLM Agents Recursive Self-Improvement+1

Sep 28, 2026cs.AI

A Persistent State for Auditable Mixture-of-Experts Routing

Mixture-of-Experts (MoE) models repeatedly route tokens to sparse subsets of experts, but conventional routers expose no routing-specific record of how cross-layer influences accumulate. We introduce Scratchpad-Augmented Mixture-of-Experts (SA-MoE), which gives each router access to a low-dimensional persistent state that is not provided to the experts. Learned layerwise writes update this state, and their realized post-update changes exactly decompose the state-mediated contribution to any later routing margin, forming a routing ledger. Across sparsely upcycled SmolLM2- and Gemma-based models and three independent training seeds per architecture, this pathway adds less than 1% analytical forward compute and is strongly used by trained routers: local removal of its router contribution changes the selected Top-2 expert set in 87.6% and 69.9% of decisions, respectively. Relative to a matched latest-write-only control, persistent accumulation increases long-horizon future-routing accessibility by 19.4 and 12.2 percentage points, with positive effects in every seed. More than 90% of absolute ledger contribution comes from non-recent writes in both families, and full-forward suppression of ledger-selected writes changes later routing and output distributions. The ledger is an exact provenance object for the persistent-state pathway, not a complete causal explanation of routing. Sensitivity-aware scores better predict full-forward intervention effects, and post-hoc methods recover related cross-layer attribution without architectural modification. SA-MoE instead makes one routing-specific computational history explicit and directly inspectable within the model's natural forward computation.

Abdurrahman Javat, Allan Kazakov

Department of Artificial Intelligence Bahçeşehir University Istanbul, Türkiye

LLM Auditing State Tracking LLM Routing Mixture-of-Experts Models

Sep 28, 2026cs.LG

FORGE: Form-Optimal Routing of Grounded Evidence for Frozen LLM Agents

In agentic AI systems, frozen foundation models are increasingly deployed as closed-weight API endpoints, making downstream adaptation possible only through the inputs and inference procedures surrounding the model. As a result, for each input query, two coupled decisions largely determine both answer quality and token cost: what evidence to provide and how much reasoning budget to allocate. Fixed defaults along these axes are often suboptimal, misallocating support form or reasoning depth on roughly 80% of queries in our analysis. To address this challenge, we propose FORGE, a unified framework for adapting frozen models through per-query routing over a joint action space that spans both support form and thinking depth. Under an entropy-regularized, cost-aware utility objective, we derive a closed-form Boltzmann routing target and instantiate the policy as a lightweight 269K-parameter factorized router. The routing policy is trained around the frozen host, without any weight access, through a three-stage pipeline: offline arm enumeration, supervised Kullback-Leibler (KL) distillation from the Boltzmann target, and Group Relative Policy Optimization (GRPO) refinement with host feedback. Across 5 knowledge-intensive benchmarks and 8 frozen backbones ranging from 7B to 671B parameters, FORGE improves accuracy at 42-45% lower token cost on both main hosts, transfers zero-shot across hosts at lower token cost, and composes with intrinsic thinking budgets where available.

Xi Xiao, Yunbei Zhang, Chen Liu +7

University of Alabama at Birmingham

LLM Grounding LLM Inference Scheduling LLM Routing

Sep 28, 2026cs.AI

Knowing When Thinking Is Not Enough: Teaching Small Reasoning Models to Reason Beyond Their Parametric Knowledge

Scaling test-time computation is a powerful way to improve language-model reasoning, and is particularly appealing for small reasoning models (sRMs) that are cheap to serve. However, is additional thinking always the right operation? By intervening at intermediate reasoning states across two model families and multiple scales, we find that self-refinement largely consolidates probability mass onto solutions already reachable from the current state, rather than making new ones reachable. These interventions reveal two failure regimes: execution bottlenecks, where the correct path is reachable and reflection can recover it, and knowledge bottlenecks, where relevant external information makes it reachable. Motivated by this distinction, we introduce FlyBy, a selective querying framework, and train 4B and 8B variants to reason first, diagnose what remains unresolved, and, at a knowledge bottleneck, query stronger models whose parametric knowledge extends beyond its own. Supervised fine-tuning bootstraps a multi-depth query action, and cost-aware reinforcement learning calibrates whether to query, what to ask, and how much to spend. On 1,158 hard problems across six benchmarks, FlyBy-4B achieves 45.96% pass@8, surpassing Qwen3-14B (41.64%) at 2.7 times lower serving cost, while also exceeding Qwen3-8B in pass@1 (16.85% vs. 15.31%). Scaling to FlyBy-8B further improves pass@8 to 51.81%.

Chanuk Lee, Minki Kang, Sangwoo Park +3

KAIST

Cost-Aware Inference Efficient Language Model Reasoning LLM Routing LLM Reasoning

Sep 28, 2026cs.LG

Routing Without Embeddings: Fast And Interpretable Routing With Regular Expressions

Large Language Model (LLM) routers commonly rely on neural query embeddings, with larger encoders expected to better capture query intent and difficulty. Yet scaling Qwen2.5 encoders from 0.5B to 72B parameters brings little improvement in routing accuracy (Figure 1b), suggesting that small encoders may already capture the query properties needed for routing. We therefore investigate which properties matter and whether they can be extracted directly from text without a neural encoder. We introduce REGEXROUTE, a pipeline that uses sparse autoencoders (SAEs) to discover interpretable regular-expression (regex) features. Using unlabeled text, an LLM turns descriptions of grouped SAE latents into regex extractors and refines them to match latent activation patterns. These extractors supply numerical features to a lightweight routing head, eliminating neural encoding at inference (Figure 1a). Across four benchmarks, one fixed set of 128 features achieves 76.43% average routing accuracy, comparable to 76.41% for the strongest neural text encoder baseline, with much smaller latency and strong robustness. These findings establish explicit, interpretable text features as a practical basis for designing and understanding LLM routers.

Yifan Lu, Qiyue Zhang, Haotian Shan +2

Rice University

LLM Inference Acceleration LLM Routing

Sep 28, 2026cs.LG

MaskCoFT: Masked Co-Adaptive Fine-Tuning for Memory-Efficient MoE Inference

Mixture-of-experts (MoE) language models often exceed the memory of a single GPU. Expert offloading keeps most experts in host memory and loads them on demand, so decoding speed depends on how many experts each token must fetch. Caching and prefetching reduce this cost only as far as the routing allows. Router-only fine-tuning can reshape the routing to reuse experts, but it keeps the experts frozen, so they cannot adapt to the tokens the new routing sends them. We propose MaskCoFT, a masked co-adaptive fine-tuning method that trains routers and experts together with the cross-entropy loss alone. During fine-tuning, a learnable binary mask restricts the Top-K routing of each layer to a subset of experts, and the experts adapt to the tokens redirected to them. At inference, the learned mask becomes a soft prior that re-ranks experts, so every expert remains selectable. We simulate a GPU cache of 4 experts per layer for Mixtral-8x7B and 12 for DeepSeek-V2-Lite. MaskCoFT cuts expert fetches per token by 23.7% and 10.1% relative to the base model. In real offloading system serving, it lowers the time per output token by up to 16.4% and 5.5%, respectively. Its average accuracy over nine benchmarks stays above the base model by 0.92 and 0.53 points.

Junfeng Wu, Zehao Fan, Hadjer Benmeziane +3

Department of Industrial and Systems Engineering, Rensselaer Polytechnic Institute · Department of Electrical, Computer, and Systems Engineering, Rensselaer Polytechnic Institute · IBM Research

Expert Offloading Mixture-of-Experts Language Models Fine-Tuning Mixture-of-Experts Inference Memory-Efficient Inference LLM Routing

Sep 24, 2026cs.CL

Baseline Shape Decides the Verdict: A Controlled Re-Examination of Ternary Language Models at 60K Parameters

Ternary (1.58-bit) weights are attractive for microcontroller-class language models, but the sub-1M-parameter regime rests mainly on isolated, single-seed comparisons. One prominent example reports that a routed ternary block (convolution, diagonal SSM and sparse attention mixed by a per-token router) beats a parameter-matched full-precision transformer by 22% at 60K parameters, attributing this to inductive bias. We re-run it under one fixed recipe, three seeds per cell, 98 byte-level runs on one laptop. (i) Baseline shape dominates: at a 16M-byte budget, param-matched transformers span 22.6% in validation loss purely by depth/width choice - far more than any architecture effect we measure there - and the best-shaped transformer ties the routed model, so the published margin is at least partly a baseline-shape effect; the ordering of shapes reverses with budget, so no single fixed shape can be trusted. (ii) At 130M bytes the routed model does win, by 22.2-24.0% over the three transformer shapes we evaluate there - but a plain gated diagonal-SSM block beats it by a further 9.1%, and the routed model's own router puts most of its weight on its recurrent pathway, so the gain does not require routing. (iii) The ternary penalty differs by architecture at the larger budget (+5.3% best transformer vs. +19.5% routed, +28.1% gated SSM), but we cannot attribute that to architecture alone: our transformers keep learned positional embeddings in full precision, 11-22% of their parameters, so they are less quantized than the models they are compared with. (iv) A 90/10 full-precision-then-ternary schedule beats all-ternary training, but only at a stage-2 learning rate about 10x the pretraining peak; at a conventional fine-tuning rate it looks 15.3% worse, reversing the conclusion. The from-scratch baseline was not itself learning-rate tuned, which bounds (iii) and (iv). Code and run logs released.

Gautam Veldanda

Independent Researcher

Ternary Quantization Quantization-Aware Training Small Language Models Byte-Level Language Model LLM Routing State Space Models

Sep 24, 2026cs.AI

CounterRoute: Self-Routed Reasoning via Hierarchical Counterfactual Credit Assignment

Reasoning-capable language models often produce long chains of thought when direct answers suffice, wasting inference compute. Many dual-mode models leave this choice to users. Automating it is challenging because routing targets evolve with the policy, initial mode preferences destabilize exploration, and sequence-level objectives entangle routing with response learning. We introduce CounterRoute, an online reinforcement-learning framework that jointly learns routing and modeconditioned responses in one shared policy directly from a native dual-mode checkpoint, without method-specific SFT warm-up. Paired current-policy counterfactual rollouts assign cross-mode credit only to the routing token, while within-mode GRPO trains response tokens. A paired-to-self-routed curriculum stabilizes early training with forced rollouts from both modes, then increases self-routed updates to improve autonomous routing. Across nine benchmarks, CounterRoute better balances accuracy and efficiency than heuristic and learned adaptive-routing methods. Relative to always-thinking checkpoints, it improves macro-average accuracy while reducing mean generated tokens by 51% for Qwen3-8B and 41% for Qwen3-14B. On instruction-following and commonsense benchmarks where direct answering is strong, think rates fall as low as 1% while response quality improves. Despite training only on math and instruction following, its routing behavior and response quality generalize to held-out coding, science, knowledge, and commonsense benchmarks.

Ruochen Jiao, Besnik Fetahu, Zhenyu Shi +1

RL for Language Model Reasoning Efficient Language Model Reasoning LLM Routing Credit Assignment Efficient Language Model Inference Counterfactual Credit Assignment+1

Sep 23, 2026cs.AI

Learning the Cost of Reliable Inference

Benchmarking and routing platforms increasingly act as intermediaries connecting large language model providers with end-users. However, providers on these platforms typically use a fixed price per token, preventing users from achieving the most competitive price for their tasks. In this work, we design a procurement platform where token prices for each task are driven by provider competition, enabling users to secure competitive pricing for guaranteed quality levels. To this end, the platform sequentially routes queries via a reverse second-price auction that incentivizes model providers to truthfully bid their best estimate of the average cost to serve a user's query. As it routes queries, the platform learns the quality offered by each provider and progressively routes queries to the most cost-competitive provider among those meeting a desired quality threshold. To validate our design, we conduct experiments with multiple LLMs from the Llama and Qwen families on popular mathematical reasoning and question-answering benchmarks. The results show that the pricing margin of the most cost-competitive provider on our platform varies significantly---from

10\%

to

71\%

---depending on the task and quality threshold. This suggests a substantial inefficiency in the current fixed-price market, and it demonstrates that our platform may enable users to capture maximum savings whenever competitive market conditions permit.

Dimitrios Rontogiannis, Ander Artola Velasco, Manuel Gomez Rodriguez

Max Planck Institute for Software Systems Kaiserslautern, Germany

Cost-Aware Inference LLM Routing

Sep 22, 2026cs.CL

COMED: The Missing Middle Between Routing and Collaboration in Multi-LLM Inference

No single Large Language Model (LLM) is uniformly reliable across queries, motivating multi-model inference systems that either route among models or combine their outputs. However, routing stops after selecting an initial model, while dense collaboration invokes peers on every query. We show that collaboration is non-monotonic: peers can recover failures that no model solves alone, but can also corrupt initially correct answers. We introduce COMED (Controlled Model Escalation for Multi-LLM Deliberation), a post-anchor controller for selective cross-model collaboration. COMED uses anchor self-consistency, router margin, and a lightweight peer probe to accept confident answers, verify ambiguous cases, and escalate only when collaboration is likely beneficial. We formalize this trade-off with a rescue-harm decomposition showing that selective collaboration improves when rescued errors outweigh collaboration-induced harms. Across medical, scientific, and general reasoning benchmarks, COMED improves fixed and routed anchors in all 16 open-weight settings, with gains up to +10.7 percentage points on MedQA while invoking fewer models and using fewer decoded tokens than dense collaboration. On HLE with frontier models, COMED improves GPT-5.5 from 23.1% to 28.1%, outperforming dense collaboration and achieving the best results.

Norah Alballa, Wenxuan Zhang, Salma Kharrat +4

KAUST · LUMS

LLM Inference Efficiency LLM Routing Multi-LLM Collaboration

Sep 16, 2026cs.CR

PentestChain: A Cost-Aware, MCP-Orchestrated Framework for Automated Penetration Testing with Free-Tier LLMs

AI-driven penetration testing has been demonstrated with premium frontier models such as GPT-4, but the per-engagement token cost makes continuous, automated testing unaffordable for the smaller organisations that need it most. This paper presents PentestChain, a ten-phase automated penetration testing framework that couples a curated, deterministic exploit map with a cost-aware AI cascade-a local Ollama model (qwen2.5-7b) first, then free-tier OpenRouter and Cerebras, with a rule-based fallback that always produces output-and exposes the full pipeline through a Model Context Protocol (MCP) server with eleven tools. We make three contributions. First, we treat US-dollar cost per engagement as a measured, first-class evaluation metric and show that a 7B-parameter local model, kept off the critical path by a deterministic backbone, sustains end-to-end operation at zero measured paid-API cost. Second, we analyse the attack surface that an MCP-exposed offensive engine introduces, grounding a four-position threat model in the 2025 MCP incident record (the CVE-2025-6514 remote-code-execution flaw in mcp-remote, the postmark-mcp supply-chain backdoor, and the tool-poisoning-rug-pull-line-jumping class), and contribute four mitigations. Third, we specify a reproducible, containerised evalua-tion protocol aligned with the standardised testbeds now expected at top-tier venues-AutoPenBench, a Cybench subset, and the PentestGPT 182-sub-task benchmark-with multi-trial statistics (more than 10 trials per configuration, pass-at-k, non-parametric significance tests and effect sizes) and direct, same testbed reproduction of the PentestGPT and PentestAgent baselines rather than citation of their published numbers. On the legacy targets measured to date, the framework detected 26 services, enriched 34 CVEs, produced

Rushabh Vipulkumar Patel, Dipo Dunsin, Mohammed Almaiah +1

Computer Security Laboratory, University of Wales Trinity Saint David, London, United Kingdom · King Abdullah IT School, The University of Jordan, Amman, Jordan · School of Computer Science and Mathematics, Keele University, Newcastle-under-Lyme, United Kingdom

Automated Penetration Testing LLMs for Cybersecurity LLM Routing

Sep 15, 2026cs.LG

Beyond the Previous Layer: Residual Predictive Structure in Sparse MoE Routing

Sparse mixture-of-experts models route each token through a sequence of expert selections. We ask whether the immediately preceding selection adequately summarizes this trajectory for predicting the next router. Using frozen OLMoE and JetMoE models, we measure the held-out predictive gain from earlier expert selections while retaining the most recent selection as a common baseline. In OLMoE, extending the history from one to eleven layers raises router-logit

R^2

from 0.59879 to 0.66544. A preregistered JetMoE replication yields four-layer gains of 0.14275 and 0.20528 at two target depths, with paired bootstrap intervals above zero. These gains survive nonlinear decoding: adding history to a small multilayer perceptron improves

R^2

by 0.17137 and 0.21861, whereas nonlinear decoding of the recent state alone adds 0.00139 and 0.00936 over a linear probe. Parameter-matched controls preserve the advantage, and cross-fitted history residuals predict target residuals with

R^2

of 0.20549 and 0.23556. These findings identify residual predictive structure in expert-selection trajectories beyond adjacent-layer persistence.

Hao Li, Yasuyuki Tahara, Yuichi Sei

Department of Informatics, Graduate School of Informatics and Engineering, The University of Electro-Communications, Tokyo, Japan

Sparse Mixture-of-Experts Expert Routing LLM Routing

Sep 15, 2026cs.IR

One Size Does Not Fit All! Dynamic Retriever and Generator Selection for RAG

Retrieval-Augmented Generation (RAG) systems typically employ fixed retriever and generator configurations across queries, despite substantial differences in query complexity and information needs, leading to inefficient allocation of computational resources. While retrieval and generation adaptivity have been studied independently, their joint effect on end-to-end RAG performance remains underexplored. We systematically analyze how retriever and generator complexity interacts across factoid and multi-hop question answering (QA), including bridge and composition reasoning tasks. Our analysis shows that stronger retrieval generally yields larger gains than increased generation effort, but both exhibit diminishing and non-monotonic returns, indicating that higher-complexity configurations are not uniformly better across queries. Motivated by these findings, we introduce DRAG, a query-adaptive framework for selecting retriever-generator configurations. We first propose DRAG

_\text{QPP}

, a training-free routing approach that uses Query Performance Prediction (QPP) signals to guide retriever selection and perplexity-based measures over retrieved context to guide generator selection. We further introduce DRAG

_\text{SFT}

, a supervised routing approach that fine-tunes an LLM to jointly predict retriever-generator configurations. Across three LLM families and four QA benchmarks, \qpprag~achieves performance comparable to strong static RAG baselines while substantially reducing inference latency, whereas DRAG

_\text{SFT}

consistently improves effectiveness over static and training-free adaptive baselines. Overall, DRAG demonstrates that jointly adapting retrieval and generation achieves a more favorable effectiveness-efficiency trade-off than static RAG pipelines.

Neeraj Anand, Payel Santra, Partha Basuchowdhuri +2

Media and Data Science Research Lab Adobe Systems, Noida, India · IACS Kolkata, India · University of Glasgow Glasgow, United Kingdom

Retrieval-Augmented Generation Adaptive RAG Question Answering LLM Routing

Sep 14, 2026cs.LG

Quality-Constrained Routing over a Fixed Pool of Quantized Mixture-of-Experts Instances

Quantized Mixture-of-Experts (MoE) services can hold several pre-materialized instances of one base model, but quantization damage varies sharply across requests and bitwidths. Because instance materialization and replica counts consume memory and require slow reconfiguration, we treat them as upstream provisioning decisions and study routing within a fixed resident pool. Within this fixed-pool boundary, we route each request to maximize modeled throughput under a class-level expected quality-degradation budget and measured instance capacities. To predict this request-specific risk, we introduce FWP (Fragility-Weighted Perplexity), computed from prompt tokens on a reference-instance prefill and calibrated to candidate-instance degradation. Underlying FWP is an exact two-expert affinity--fragility decomposition and a conditional multi-layer top-

k

expansion whose bias, interaction, route-change, separability, and higher-order terms remain explicit. Using these calibrated risks, a window-level linear program yields a signed reduced-reward score that is KKT-consistent with the LP optimum under optimal prices and primal-feasible tie allocation. On 88 extended Qwen prompts, complete W2, W3, and W4 instances quantizing all 6,144 expert blocks incur mean

\Delta

NLL of

0.9437

,

0.1832

, and

0.0513

. Under the same population and

\tau=0.1513

, FWP allocation reaches a

1.284\times

offline model-based multiplier versus

1.253\times

for request-agnostic mixing and

1.000\times

for static W4, an incremental

2.5\%

relative FWP gain.

Zhenghong Huang, Hongfan Wu, Jiheng Zhang

The Hong Kong University of Science and Technology Clear Water Bay, Kowloon, Hong Kong

Mixture-of-Experts Inference Mixture-of-Experts Quantization Constrained Optimization LLM Routing