---
schema_version: "newruntime-topic-hub-v0.1"
type: "topic_hub"
slug: "model-routing"
title: "Model routing - New Runtime"
description: "How teams route work across model families, effort levels, cheap executors, judges, and local/runtime constraints."
answer: ["Model routing treats models as swappable execution resources rather than a single product choice.","The pattern is to match task risk, latency, cost, and verification needs to the right model or model sequence.","This makes the harness more durable than any one frontier model release."]
search_intents: ["model routing","multi model AI systems","LLM routing architecture"]
status: "featured"
last_updated: "2026-07-16"
counts: {"total":13,"signals":11,"patterns":2,"posts":0,"atlas":0,"sources":25}
routes: {"html":"https://newruntime.com/topics/model-routing/","markdown":"https://newruntime.com/topics/model-routing.md","json":"https://newruntime.com/topics/model-routing.json"}
top_sources: ["https://claude.com/blog/claude-model-and-effort-level-in-claude-code","https://code.claude.com/docs/en/model-config","https://cognition.com/blog/making-fable-cheaper-than-opus","https://databricks.com/blog/benchmarking-coding-agents-databricks-multi-million-line-codebase","https://docs.inference.net/introduction","https://engineering.ramp.com/post/ai-spend-value","https://github.com/langchain-ai/deepagents","https://github.com/NousResearch/hermes-agent","https://github.com/rivet-dev/agentos","https://github.com/shadcn/improve","https://github.com/shadcn/improve/blob/main/skills/improve/SKILL.md","https://github.com/steipete/agent-scripts/blob/main/skills/codex-first/SKILL.md"]
---

# Model routing - New Runtime

How teams route work across model families, effort levels, cheap executors, judges, and local/runtime constraints.

## Short Answer

- Model routing treats models as swappable execution resources rather than a single product choice.
- The pattern is to match task risk, latency, cost, and verification needs to the right model or model sequence.
- This makes the harness more durable than any one frontier model release.

## Patterns

- [Harness architecture outlives model choice](https://newruntime.com/patterns/harness-architecture-outlives-model-choice/): For production agents, the harness is becoming a more durable product boundary than the identity of the model running inside it.
- [Agent economics moves to completed work](https://newruntime.com/patterns/agent-economics-moves-to-completed-work/): The economically meaningful unit for agent systems is becoming cost per verified completed task rather than cost per token or model call.

## Field Notes


## Recent Raw Signals

- 2026-07-16: [OpenCode makes the model a swappable dependency](https://newruntime.com/signals/opencode-model-swappable-agent-harness/)
- 2026-07-15: [A pricier model can be cheaper per completed task](https://newruntime.com/signals/fable-cost-per-completed-task/)
- 2026-07-13: [Claude Code separates model choice from effort](https://newruntime.com/signals/claude-code-model-and-effort-are-separate-controls/)
- 2026-07-12: [Improve routes architecture and execution to different models](https://newruntime.com/signals/improve-routes-architect-and-executor-models/)
- 2026-07-09: [An advisor model can guide a cheaper executor](https://newruntime.com/signals/advisor-model-guides-cheaper-executor/)
- 2026-07-02: [An inference gateway makes model switching operational](https://newruntime.com/signals/inference-gateway-makes-model-switching-operational/)
- 2026-06-28: [Coinbase Reports Lower AI Spend as Token Use Rises](https://newruntime.com/signals/coinbase-reports-lower-ai-spend-as-token-use-rises/)
- 2026-06-27: [GLM-5.2 Changes the Economics of Claude Code Workflows](https://newruntime.com/signals/glm-5-2-changes-the-economics-of-claude-code-workflows/)
- 2026-06-25: [Sakana Fugu Presents Multiple Agents as One Model Endpoint](https://newruntime.com/signals/sakana-fugu-presents-multiple-agents-as-one-model-endpoint/)
- 2026-06-17: [Fusion API Uses Multiple Models and a Judge for One Answer](https://newruntime.com/signals/fusion-api-uses-multiple-models-and-a-judge-for-one-answer/)
- 2026-05-15: [Vercel Treats Production AI as Multi-Model Infrastructure](https://newruntime.com/signals/vercel-treats-production-ai-as-multi-model-infrastructure/)

## Public Sources

- https://claude.com/blog/claude-model-and-effort-level-in-claude-code
- https://code.claude.com/docs/en/model-config
- https://cognition.com/blog/making-fable-cheaper-than-opus
- https://databricks.com/blog/benchmarking-coding-agents-databricks-multi-million-line-codebase
- https://docs.inference.net/introduction
- https://engineering.ramp.com/post/ai-spend-value
- https://github.com/langchain-ai/deepagents
- https://github.com/NousResearch/hermes-agent
- https://github.com/rivet-dev/agentos
- https://github.com/shadcn/improve
- https://github.com/shadcn/improve/blob/main/skills/improve/SKILL.md
- https://github.com/steipete/agent-scripts/blob/main/skills/codex-first/SKILL.md