If you are lucky and there are no significant interaction effects between brand and price (check on that, first by looking at your 2-way counts analysis and seeing if the Interaction Chi-Square is significant), then you just analyze is (almost) the same as you would a traditional conjoint design.
You will get 3 utility values for brand, and also 3 utility values for price, assuming you continue with a main-effects model specification.
If you want to make a prediction regarding how respondents like the first brand at its first level of price (whatever that price was, specific to that brand), you add the brand utility to the price utility (for the first level of price). Of course, the first level of price is a different price level for brand, but that doesn't matter to your predictions.
One key thing to remember is that you cannot interpret your brand utilities as the independent worth of that brand, holding all other attributes equal...because you did not hold all other attributes equal. Your brand utilities are the average utility for brand when it was shown at its average price. But, that should matter to you for the purposes of making market simulator predictions.
When you are using market simulators to predict shares of choice under different competitive scenarios, you are not trying to interpret the utility coefficients. You are just seeing how combinations of attribute levels lead to different shares of preference under different competitive scenarios, so the interpretation and use of the conjoint simulator is the same as a standard conjoint study.