adaptive-routing

Stop overpaying to run your agents. Kalibr routes every request to lower-cost model and tool paths without degrading performance.

Offloading DOCA-based Adaptive Routing onto NVIDIA BlueField-2 DPU. (卸载基于DOCA的自适应路由到NVIDIA BlueField-2 DPU上)

Official implementation of MAGNET: Modality-Guided Mixture of Structured Experts with Entropy-Triggered Routing for Multimodal Recommendation.

Krako 2.0 – Energy-efficient, triadic multi-tier inference infrastructure enabling adaptive routing across heterogeneous edge–cloud nodes.

Multi-model AI gateway with protocol conversion, adaptive routing, billing, and diagnostics.

Adaptive job router for Tokio: per job, choose inline, spawn, CPU pool, batch or drop from job cost, latency budget and live pressure, with online learning, Prometheus metrics and a live dashboard.

VPN-клиент, который сам выбирает маршрут по качеству и способу обхода, а не по пингу. Hysteria2, VLESS+Reality, Trojan, Shadowsocks-2022.

Adaptive per-turn AI model routing for Agy, Codex, and Claude — automatically choose the right model tier while preserving each provider's native workflow.

Ant Colony Optimization algorithm for the shortest path problem

A self-organizing distributed intelligence network for adaptive routing, resilient communication, peer discovery, failure recovery, and decentralized computation.

Learning Automata algorithm for the shortest path problem

Hermes ARC (Adaptive Routing Core) — adaptive conversational orchestration plugin for Hermes Agent

Particle Swarm Optimization algorithm for the shortest path problem

Reproduction and extension of TISER for temporal question answering, using Qwen2.5-3B-Instruct with LoRA and an adaptive TF-IDF-based router to balance reasoning accuracy and generation cost.

LLM Router Gateway and Adaptive Router

Qomni Planner v0.59: Adaptive Cognitive Routing with Online Learning Loops for Edge AI Inference — Percy Rojas Masgo, CEO Condesi Perú · Qomni AI Lab

Protocol-aware AI model & agent gateway with feasibility-first routing, failover, OpenAI/Anthropic/Responses compatibility, deterministic simulation, adaptive routing research, and reproducible cost/latency evidence.

Adaptive LLM router that learns which route to use on every request.

KERNEL v4.2 est un scheduler cognitif adaptatif pour LLMs intégrant LATS, Reflexion et Meta-Prompting. Il optimise dynamiquement le routing du raisonnement selon la complexité (1-5) pour maximiser la précision et réduire les hallucinations. Propulsé par la vision Lichen-Collectives.

LLM routing engine that dispatches prompts to the cheapest capable model