IQ Routing

The product is a gateway that sits in front of LLM calls for chatbots, RAG pipelines, agent loops and finance workloads and routes each request to the cheapest model that meets a quality bar, addressing overspending from pinning every step to a top-tier reasoning model and lack of cost attribution. It sells to developers, data and operations teams in B2B companies building AI applications. It is delivered as a SaaS and infrastructure API that drops in behind a single base URL in front of OpenAI or Anthropic SDKs.

Key features

• Per-request routing to cheapest qualifying model
• Capability aliases for intent-based model selection
• Per-step thinking budget resolution
• Session-aware routing within model families
• Semantic and exact response caching
• Per-team budgets and per-key limits
• Four-way cost split ledger
• Fallback on provider failure
• Complexity-based classifier for request difficulty