If you’ve ever kicked off a complex task in Copilot and thought, “Why is it using that model?”, you’re not imagining things. This piece digs into why Copilot’s Auto mode isn’t actually choosing models based on what you’re asking it to do.
The short version: Auto is optimized for availability, quotas, and keeping the service running smoothly, not for matching model brainpower to task complexity. There’s even a small usage discount baked in if you stick with Auto, which adds another wrinkle.
It’s a great look behind the curtain at how Copilot really routes requests today, and why manual model selection still matters if you care about consistent reasoning.
Full article here:
Why Copilot’s Auto Mode for AI Models Ignores Your Actual Task - Visual Studio Magazine
Do you trust Auto, or do you always pick your model yourself?