A post by @chetaslua says Microsoft's EF Core team merged a change on October 16 that sets gpt-6.1-terra as the default judge model in 18 evaluation configs. The previous default was gpt-5.6-terra. The post says the calls run through GitHub Copilot's SDK. The post includes an image.

Image shared by @chetaslua on X
Image shared by @chetaslua on X

Image credit: @chetaslua on X

What the change does

A judge model is the model that grades another model's outputs in an evaluation run. Changing the default judge changes the scores those evals produce, even when the models being tested stay the same. That is why a switch in 18 configs matters to anyone comparing results across repos or time.

What readers can check

If you maintain EF Core evaluations or copy its configs, look for judge_model settings and confirm which model they call. Scores from before and after such a change are not directly comparable.

Sources