Tag
Routed Graph Handoff proposes a lightweight router to adaptively select between structured graphs and natural language for inter-agent communication in multi-agent LLM systems, improving benchmark performance while reducing token cost.
RetrievalRouter is a lightweight query-aware router that adaptively selects retrieval pipelines to improve accuracy and speed in document retrieval, outperforming static baselines.
Cognition introduces Devin Fusion, an adaptive model router that reduces cost by 35% while maintaining real frontier intelligence for agentic coding tasks.
Discusses the shift from token-based productivity metrics to output, impact, and value measurement in AI adoption, highlighting Cognition's solutions: adaptive routing, spend attribution, automations, and a productivity guarantee.
AgentCodec is a source-available library unifying 28 LLM reliability techniques (retries, ensembling, generator/critic refinement, etc.) under a single OpenAI-compatible API, with adaptive routers that can reduce inference costs by ~56% at matched quality. It adopts a communication-theory framing and supports drop-in replacement for OpenAI, Anthropic, and Ollama clients.