HyDRA: Hybrid Dynamic Routing Architecture for Heterogeneous LLM Pools

• URL: http://arxiv.org/abs/2605.17106v1
• Date: Sat, 16 May 2026 18:19:30 GMT
• Title: HyDRA: Hybrid Dynamic Routing Architecture for Heterogeneous LLM Pools
• Abstract summary: We present HyDRA, a framework that predicts fine-grained, multi-dimensional capability requirements per query.<n>A shortfall-matching algorithm then selects the cheapest model whose capabilities meet the predicted requirements.<n>The framework is deployed to all users in GitHub Copilot's VS Code Chat auto-mode.
• Score: 2.0320563270126493
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Production LLM deployments increasingly maintain heterogeneous model pools spanning order-of-magnitude cost differences. Existing routers make binary strong-vs-weak decisions and couple learned parameters to specific model identities, requiring retraining whenever the catalog changes. We present HyDRA (Hybrid Dynamic Routing Architecture), a framework that predicts fine-grained, multi-dimensional capability requirements per query and matches them against configuration-defined model profiles via shortfall matching. A ModernBERT encoder with K=4 independent sigmoid heads scores each query along reasoning, code generation, debugging, and tool use; a shortfall-matching algorithm then selects the cheapest model whose capabilities meet the predicted requirements. The deployed predictor runs at 86 ms median CPU inference latency in production, and is fully decoupled from the model catalog -- adding or removing models requires only a configuration change, with zero retraining. On SWE-Bench Verified (5-model pool: GPT-5.4-mini, Claude Haiku 4.5, GPT-5.3 Codex, Claude Sonnet 4.6, GPT-5.4), HyDRA's tunable shortfall threshold spans three regimes: peak-quality exceeds the always-strong Claude Sonnet 4.6 baseline (75.4% vs. 74.2% resolution) at 12.9% cost savings; iso-quality matches Sonnet at 54.1% cost savings, a 6x improvement over our prior in-house binary router at 9.1%; aggressive pushes savings to 72.5% for a 3.2-point quality trade. Results generalize across LiveCodeBench, BigCodeBench, and tau-bench. HyDRA is deployed to all users in GitHub Copilot's VS Code Chat auto-mode and -- to our knowledge for the first time in the LLM routing literature -- demonstrates language-invariant routing across CJK, European, and other script families.

Related papers

• From Production Traffic to Post-Training: Building a Self-Hosted LLM That Covers the Corporate Request Mix [55.08143827481755]
  We consolidate traffic from over 200 internal applications onto a single model.<n>We train a separate GRPO expert per axis and merge them via two-stage SLERP.<n>The model absorbs 50% of platform traffic, 116M requests per month, at a fraction of the serving cost.
  arXiv  Detail & Related papers  (2026-09-01T17:39:26Z)
• On the Design of Qwen3.8-Next Architecture: Evaluation, Efficiency, and Training Stability [57.03393882933515]
  We describe the architecture and ablations of Qwen3.8-Flash-Next, a sparse mixture-of-experts model with 125B parameters, 6B activated per token, and 51B parameters of n-gram embedding tables held off the accelerator.<n>On fourteen pre-training benchmarks the model leads the 397B-A17B predecessor on eight and trails it on the rest by at most 2.6 points, at 1/3 the activated parameters, 1/3 the training tokens, and roughly 1/9 the training FLOPs.
  arXiv  Detail & Related papers  (2026-08-31T06:35:07Z)
• Architecture as Capability Equalizer for Coding Agents [0.8741110159894001]
  LLM-based coding agents generate complete software systems from high-level descriptions.<n>We present a controlled experiment comparing five informationally equivalent specification formats.<n>Across 90 multi-turn agent trials, specification format shows a strong format x model interaction.
  arXiv  Detail & Related papers  (2026-08-22T03:04:12Z)
• VisualClaw: A Real-Time, Personalized Agent for the Physical World [76.89211120690028]
  We present VisualClaw, a self-evolving multimodal agent built around two principles.<n>First, hybrid encoding reduces deployment cost by filtering less informative streaming frames with a cascaded gate.<n>Second, skill evolution lets the agent learn from failures, producing skill-bank updates that help future questions.
  arXiv  Detail & Related papers  (2026-06-15T06:58:22Z)
• Distributional Energy-Based Models for Uncertainty-Aware Structured LLM Reasoning [40.342912574072024]
  Large Language Models produce structured outputs such as travel plans or code solutions.<n>Individual reasoning steps may appear correct while the output as a whole violates budgets, fails test cases, or contradicts earlier deductions.<n>We propose a learned quality scorer with deterministic analytical constraint penalties for verifying structured LLM outputs.
  arXiv  Detail & Related papers  (2026-05-15T17:08:27Z)
• ChipMATE: Multi-Agent Training via Reinforcement Learning for Enhanced RTL Generation [55.947962672433675]
  ChipMATE is the first self-trained multi-agent framework for RTL generation.<n>Inspired by industrial practice, ChipMATE pairs a Verilog agent with a Python reference-model agent that mutually verify each other's outputs.<n>ChipMATE achieves 75.0% and 80.1% pass@1 on VerilogEval V2 with 4B and 9B base models.
  arXiv  Detail & Related papers  (2026-05-13T01:04:21Z)
• When Correct Isn't Usable: Improving Structured Output Reliability in Small Language Models [2.064923532131528]
  Deployed language models must produce outputs that are both correct and format-compliant.<n>We study this structured-output reliability gap using two mathematical benchmarks -- GSM8K and MATH.<n>We develop AloLab, an iterative system-prompt requiring only black-box API access to the target model.
  arXiv  Detail & Related papers  (2026-05-04T09:07:44Z)
• SAGAI-MID: A Generative AI-Driven Middleware for Dynamic Runtime Interoperability [1.2744523252873352]
  SAGAI-MID is a FastAPI that uses large language models (LLMs) to dynamically detect and resolve schema mismatches at runtime.<n>We evaluate SAGAI-MID on 10 interoperability scenarios spanning version migration, IoT-to-analytics bridging, and protocol conversion across six LLMs from two providers.
  arXiv  Detail & Related papers  (2026-03-30T17:46:41Z)
• KAT-Coder-V2 Technical Report [45.53423367042574]
  We present KAT-Coder-V2, an agentic coding model developed by the KwaiKAT team at Kuaishou.<n>KAT-Coder-V2 adopts a "Specialize-then-Unify" paradigm that decomposes agentic coding into five expert domains.<n>We develop KwaiEnv, a modular infrastructure sustaining tens of thousands of concurrent sandbox instances.
  arXiv  Detail & Related papers  (2026-03-29T14:01:35Z)
• Resource-Efficient Iterative LLM-Based NAS with Feedback Memory [49.44875022114861]
  Neural Architecture Search (NAS) automates network design, but conventional methods demand substantial computational resources.<n>We propose a closed-loop pipeline leveraging large language models (LLMs) to iteratively generate, evaluate, and refine convolutional neural network architectures.
  arXiv  Detail & Related papers  (2026-03-12T16:00:22Z)
• Efficient Multi-Model Orchestration for Self-Hosted Large Language Models [2.3275796286410677]
  Pick and Spin is a framework that makes self-hosted orchestration and economical.<n>It integrates a unified Helm-based deployment system, adaptive scale-to-zero automation, and a hybrid routing module.<n>It achieves up to 21.6% higher success rates, 30% lower latency, and 33% lower cost per query compared with static deployments of the same models.
  arXiv  Detail & Related papers  (2025-12-26T22:42:40Z)
• Towards a Science of Scaling Agent Systems [79.64446272302287]
  We formalize a definition for agent evaluation and characterize scaling laws as the interplay between agent quantity, coordination structure, modelic, and task properties.<n>We derive a predictive model using coordination metrics, that cross-validated R2=0, enabling prediction on unseen task domains.<n>We identify three effects: (1) a tool-coordination trade-off: under fixed computational budgets, tool-heavy tasks suffer disproportionately from multi-agent overhead, and (2) a capability saturation: coordination yields diminishing or negative returns once single-agent baselines exceed 45%.
  arXiv  Detail & Related papers  (2025-12-09T06:52:21Z)
• GLM-4.5: Agentic, Reasoning, and Coding (ARC) Foundation Models [194.64264251080454]
  We present GLM-4.5, an open-source Mixture-of-Experts (MoE) large language model with 355B total parameters and 32B activated parameters.<n>Through multi-stage training on 23T tokens and comprehensive post-training with expert model iteration and reinforcement learning, GLM-4.5 achieves strong performance across agentic, reasoning, and coding tasks.<n>We release both GLM-4.5 (355B parameters) and a compact version, GLM-4.5-Air (106B parameters), to advance research in reasoning and agentic AI systems.
  arXiv  Detail & Related papers  (2025-08-08T17:21:06Z)
• I Know Which LLM Wrote Your Code Last Summer: LLM generated Code Stylometry for Authorship Attribution [0.0580448704422069]
  This paper presents the first systematic study of authorship attribution for C programs.<n>We release CodeT5-Authorship, a novel model that uses only the encoder layers from the original CodeT5 encoder-decoder architecture.<n>Our model achieves 97.56% accuracy in distinguishing C programs generated by closely related models.
  arXiv  Detail & Related papers  (2025-06-18T19:49:41Z)
• S*: Test Time Scaling for Code Generation [55.11863577956177]
  We propose S*, the first hybrid test-time scaling framework for code generation.<n>S* substantially improves the coverage and selection accuracy of generated code.
  arXiv  Detail & Related papers  (2025-02-20T09:18:53Z)
• Hybrid-Segmentor: A Hybrid Approach to Automated Fine-Grained Crack Segmentation in Civil Infrastructure [52.2025114590481]
  We introduce Hybrid-Segmentor, an encoder-decoder based approach that is capable of extracting both fine-grained local and global crack features. This allows the model to improve its generalization capabilities in distinguish various type of shapes, surfaces and sizes of cracks. The proposed model outperforms existing benchmark models across 5 quantitative metrics (accuracy 0.971, precision 0.804, recall 0.744, F1-score 0.770, and IoU score 0.630), achieving state-of-the-art status.
  arXiv  Detail & Related papers  (2024-09-04T16:47:16Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.