# Hugging Face — Model Routing Is Simple. Until It Isn’t.

- Company: Hugging Face (huggingface.co)
- Announced: 2026-07-15T17:27:01+00:00
- Subject: Platform
- Models affected: GPT-4.1, Claude Sonnet 4.6, Opus
- Source: https://huggingface.co/blog/ibm-research/model-routing-is-simple-until-it-isnt
- Record: https://forck.live/items/1456-model-routing-is-simple-until-it-isn-t

The article discusses challenges in model routing for agentic systems, showing that actual cost depends on caching, difficulty is often invisible at routing time, and latency is affected by infrastructure. The authors propose treating routing as an optimization problem rather than classification, and present results on AppWorld Test Challenge showing a cost-accuracy frontier with configurations achieving 84% accuracy at $93 and 83s latency, a 21% cost reduction and 9% latency reduction compared to running Opus alone with only a 4% accuracy drop.

## Evidence

Verbatim from https://huggingface.co/blog/ibm-research/model-routing-is-simple-until-it-isnt:

> The lesson we took away from this work is that routing isn’t really about choosing models. It’s about optimizing systems.

---

Record: https://forck.live/items/1456-model-routing-is-simple-until-it-isn-t
Catalogue: https://forck.live/llms.txt
Feed: https://forck.live/feed.md