IQ Routing – route every call to the cheapest model that holds quality

IQ Routing, a new AI gateway, claims to cut LLM call costs by 40 to 80 percent by routing each request to the cheapest model that meets a quality bar, with a five-step LangChain loop example dropping from $1.84 to $0.77 (58 percent less). The service drops in front of OpenAI or Anthropic SDKs, supports Claude Code and Cursor, and offers features like caching, per-team budgets, and session governance.

Route every call to the cheapest model that holds quality. IQ Routing drops in front of chatbots, RAG pipelines, agent loops, and finance workloads. Cut spend 40 to 80 percent, measured on our own traffic, live in thirty seconds. Card required. No charge on Free. Drop in front of the OpenAI or Anthropic SDK you already use Any OpenAI- or Anthropic-shaped endpoint drops in behind one base URL. Kimi and GLM are next on the roadmap. The problem Every call hits your top model. The bill keeps climbing and nobody can say why. Pin one frontier model on classification, retrieval, tool calls, and cleanup alike and you pay reasoning rates for work a cheap model would nail. Then the invoice lands as one line, shipping outruns anything finance can categorise, the cost report is a guess, and the audit trail is a Slack thread. One worked example, before and after The same five-step LangChain loop, routed two ways. The Before trace pins every step on the same top-tier reasoning model, which totals $1.84. The After trace lets the resolver pick the cheapest variant that holds the quality bar on each step, which totals $0.77, or 58 percent less for the same five steps. Every per-step price is on the page, so the arithmetic is yours to check rather than ours to assert, and nothing in the agent code changes. Before · vanilla LangChain one provider, one budget Every step on claude-opus-4-8 at thinking=high - Planningclaude-opus-4-8 · thinking=high$0.42 - Retrievalclaude-opus-4-8 · thinking=high$0.38 - Tool callclaude-opus-4-8 · thinking=high$0.28 - Synthesisclaude-opus-4-8 · thinking=high$0.41 - Verificationclaude-opus-4-8 · thinking=high$0.35 After · IQ Routing per-step resolver, session envelope The resolver picks the cheapest variant per step - Planningcap:reason-heavy · gpt-5.6-terra high$0.25 - Retrievalcap:long-context-128k · gpt-4o$0.06 - Tool callcap:tool-call-strict · gpt-4o no thinking$0.04 - Synthesiscap:reason-heavy · claude-opus-4-8 high$0.41 - Verificationcap:cheap-fast · gpt-4o-mini$0.01 The fix One gateway. Six superpowers. Each one shows up on the dashboard the moment your first request lands. State /docs/api-reference Works inside the apps that lock models. Some tools only allow one model family. Point Claude Code, Cursor, or your own app at IQ and it stays inside the family they require, routing to the right tier for each step and never leaking to another vendor. It holds the conversation too, so switching models mid-thread stays safe. Read the docs → Cache /docs/caching Repeats resolve before they bill. When a request matches one IQ has already handled, the answer comes straight back and you are not charged for it again. IQ catches exact repeats and the ones that just mean the same thing, and every team's cache stays walled off from the rest. Read the docs → Route /docs/routing The router is the hard part. Routing well is the whole product. A purpose-built classifier reads each request for true difficulty, weighs it against live cost and latency across providers, and picks the model that clears your quality bar for the least spend, with instant fallback if one slips. It holds quality where a naive cheapest-model router quietly drops it. Read the docs → Book /docs/billing Every dollar has an owner. Per-team budgets, per-key limits, and a four-way input, output, cached, and thinking cost split, written to the ledger and enforced before the provider call. Read the docs → Session /docs/sessions A session is one accountable unit. Every multi-step agent loop tracks per-turn cost, latency, tokens, and step class, and a governor caps how much thinking the session can spend. Read the docs → Step /docs/step-classifier Each step gets the right model. An agent runs many steps, and most of them don't need your most expensive model. IQ sizes up each one and routes it to the cheapest model that can still do it well, so you get frontier quality where it counts and pay a fraction everywhere else. Read the docs → What the dashboard looks like Spend, budgets, and per-team usage on one page. Illustrative dashboard, sample data. Pricing Priced so it pays for itself. Start free in thirty seconds. Scale into Team and Enterprise when the bill makes the case. Free $0 forever. Card on file, no trial, no charge to start. One URL for every provider. Bring your own provider keys BYOK . - ✓OpenAI, Anthropic, and Google, with one URL for all three - ✓Bring your own keys BYOK - ✓Semantic cache - ✓Email support Start free /sign-up?plan=free Team Most popularBilled monthly, no annual plan Everything in Free plus alerting, org-scoped access, and custom band maps. - ✓Everything in Free - ✓Per-team budgets - ✓Audit log - ✓Per-team alerting - ✓Org-scoped access controls - ✓Custom band maps auto, cheap, frontier - ✓Priority email support Upgrade to Team /pricing?plan=team Enterprise Custom annual contract On-prem deployment, SOC2 evidence pack on request, dedicated support. - ✓Everything in Team - ✓On-prem or VPC deployment - ✓SOC2 evidence pack on request - ✓ERP integrations on the roadmap - ✓Dedicated support engineer Contact sales mailto:support@iq-routing.com?subject=IQ%20Routing%20Enterprise Three lines today. A clean bill tomorrow. Route real traffic in under five minutes. Card required. No charge on Free.