Choosing the right statistical test?

Hello! I'm trying to choose the best statistical test! Suppose I work for a car company ("Company A") that designed a new car (call it "car A"). I have data about "car A"'s MPG (miles per gallon) from 200 different car As. I also have essentially the same exact data for a competitor company's similar version of the car (call it car "B") with 200 numbers for that car's MPG. Now, I just want to compare car A's MPG to car B's MPG. I'll use a random sample of 30 rows from each of the 200 MPG populations to compare the two car's fuel efficiency.

Which would be the best statistical test to choose for determining whether car A's mpg is statistically different from car B's mpg?

These are my statistical test choices:
1. One-sample t-test
2. Two-sample t-test (Unpaired)
3. Two sample t-test (Paired)
5. Chi-Squared

To clarify, I think I'm just wanting to compare two samples from the two different populations to determine if they are statistically different. Hope that makes sense. I will GREATLY appreciate ANY insights on this! Thanks SO much in advance!