Did the Draw Go to Plan?

A grocery delivery app A/B tests a new recommender. The plan is that a random draw sends 1 in 4 customers to the challenger, and everyone else keeps the champion. After two weeks the dashboard shows:

Before comparing order rates, the team runs a sample ratio mismatch (SRM) check. It asks whether the draw really split customers in the planned ratio. If something broke the split, such as a bug that drops some challenger customers, then the groups may no longer be alike, and comparing their outcomes can't be trusted.

The check uses the chi-square goodness-of-fit statistic:

$$\chi^2 = \sum_{\text{groups}} \frac{(O - E)^2}{E}$$

where $O$ is a group's observed count and $E$ is the count the planned split predicts for that group, given the total.

For two groups, a $\chi^2$ above $10.83$ would turn up less than 1 time in 1,000 if the draw were working as planned. Many teams treat that as a broken experiment.

What is $\chi^2$ for this test? Round your answer to 2 decimal places.