We built a model router that plugs into coding agents (e.g. Claude Code, Codex, Cursor, etc.) and intelligently sends requests to the best model to serve them. Here's a quick demo of running it locally: https://youtu.be/isKhAyivtfM

At Weave, we write ~all our code with AI, and it's been getting more expensive. This came to a head when Opus 4.7 was released and, thanks to its tokenizer changes, our costs shot up. We knew we didn't need Opus for everything but we didn't want to lose out on the intelligence for the cases where you really need it. So we decided to build a model router to handle this for us.

The Weave Router acts as an Anthropic/OpenAI endpoint specifically for coding agents. It looks at every inference request and intelligently (more on that in a sec) decides what model to send it to, handling all the translations required along the way. So it can use faster/cheaper models (e.g. DeepSeek v4, GLM 5.2, Kimi K2.6) when possible, and frontier models (Opus 4.8 & GPT 5.5 (& Fable whenever it's back)) when necessary.

How do we know what model to route to? We trained an RL model on tens of thousands (so far!) of agent traces. We reward the routing model when it selects an LLM that successfully completes the given task.

Here's an example: if you ask the router to plan a complex change, it will (probably) route that request to Opus 4.8. Subagents exploring the codebase to gather context will be routed to more suitable models (e.g. DeepSeek V4 Flash). Then when you have the plan ready to implement, it will be (most likely) be handed to a quicker model (e.g. GLM 5.2) to carry it out.

We've been using this internally for the last month or so. We've saved 40% on tokens vs. what we otherwise would have paid, with no noticeable differences in quality or velocity.

The router is source-available under Elastic License 2.0, so you can self-host it. Or if you prefer, you can also use our hosted version: weaverouter.com.

I'll be here to answer any questions you may have!

Combining LLMs Rarely Beats the Best Single Model, I tested 67 frontier models

SQLite: Past, Present, and Future (2022)

Testing Grok Imagine's 15-20x Faster Image Generation

Show HN: Smart model routing directly in Claude, Codex and Cursor

A free checker for whether AI search engines can cite your site

1.38 Millimeter Microcontroller

An eBook dictionary generator for Dungeon Crawler Carl, or any other fandom

GLP-1 drugs increase women's marriage and employment rates

AI Cheerleading, AI Abstention and AI Redirection

Samsung will soon start charging to access its smart home API

Stop Slopping Your Skills

Europe to exclude military-age Ukrainian men from temporary protection scheme

Show HN: AI Is Too Good at Coding, Let's Make It Better

OpenAI set to limit GPT 5.6 rollout after call from Trump administration

Silk is a stackful-fiber library, scheduler with a work-stealing loop, io_uring

The Coming Divide: AI-Native or Left Behind

Getting AI credits for reading ads while your coding agent is doing its thing

The road expands before me

Engineer accused of insider trading around Microsoft's Three Mile Island reboot

A Dark Dimension Could Link Two of the Universe's Great Unknowns

Everything Is AI

I built a app to help founders plan and build organized apps – AMA

AI and Liability

One man, two kernels, and a lot of RISC-V

Un0rick: Open-source pulse echo ultrasound

Show HN: Uv-matrix – A tiny matrix runner for Python projects using Astral uv

Best Investments over the Last 100 Years? Almost All Are Tech Companies

Venetian Bridge Brawls in 17th and 18th Century Art

Alamo Drafthouse spent years banning phones. Now it requires them

Show HN: OpenOrb – BYOK multi-AI council that shows its work

Show HN: Apply for jobs by directly emailing relevant people at a company