Firefox used ~50% less memory than Chrome on open pages, but the composite memory score was slightly for Chrome because after closing the tabs Firefox used more memory over its starting amount. Tom's basically treated this as a leak when really it was just fixed overhead (wouldn't grow when opening and closing the tabs over and over).
FF 'won' this benchmark despite a unfair memory score against it.
FF 'won' this benchmark despite a unfair memory score against it.