Drift Detection for LLM Routing: Catching Silent Model Degradation

<p>Your AI system isn't going to crash. It's going to drift.</p> <p>A recommendation engine making 1.4 model calls instead of 1. A retrieval pipeline fetching 5 chunks instead of

<p><a class="article-body-image-wrapper"

<p><em>Disclosure: I maintain <a href="https://github.com/Fast-Editor/Lynkr" rel="noopener noreferrer">Lynkr</a>, an open-source router whose design decisions this post explains.

<!-- SC_OFF --><div class="md"><p><strong>Submitted by:</strong> Adam Kruger <strong>Date:</strong> March 23, 2026 <strong>Models Solved:</strong> 3/3 (M1, M2, M3) + Warmup</p>

<p>I run a public board that probes 16 LLMs on a frozen 35-task suite, once a day, and keeps every score. When a model drops against its previous run, it opens a GitHub issue by

<p>Anthropic shipped Opus 4.7 on April 16, 2026. Same $5/$25 pricing as 4.6. Same 1M context. Same 128K output ceiling. Different model entirely when you put it in an agentic