The Unbiased Model That Loses

Two models predict the same target. On new data their expected squared error decomposes as

$\text{Expected test error} = \text{Bias}^2 + \text{Variance} + \sigma^2_{\text{noise}}$

Measured on a held-out set:

The irreducible noise variance is $\sigma^2_{\text{noise}} = 5$ for both models.

Which model has the lower expected test error, and what is that error?