We're taking our time with the new MacBook Airs, trying to find out if you can really count on these wonder wedges to serve as your main machine. But some early benchmarks are already out, and here's how they look:
MacWorld ran a thorough Speedmark 6.5 benchmark on all four standard new MacBook Air models.
Surprisingly, MacWorld found that the new 1.86GHz MacBook Airs trounced the older Airs across the board, grabbing an overall score of 108 compared to the older 1.86GHz MBA's 54 and the 2.13GHz MBA's 63 . The solid state drives helped the new MBAs scorch their predecessors in drive-intensive tests, and even processor-heavy tests involving iTunes encoding and Photoshop were snappier on the new Airs, despite their having the same 1.86GHz processor. That slower clock speed also kept the new MBAs from getting overly warm, a problem that was common with older versions.
Less surprisingly is that the 13" MacBook Air outperformed the Core2Duo 13" MacBook Pro in many tests, edging it out overall with a score of 108 compared to the MBP's 106. The new MBAs were 59% faster than the 13" MacBook Pro in file duplication tests, 43% faster in file unzipping, and roughly 21% faster in MacWorld's Call of Duty test.
When it came to video encoding, though, the 13" MacBook Pro won out: it was 20% faster than the new 13" Air and 43% faster than the 11". And while the new MacBook Air's flash storage kept it apace with the 15" Core i5 MacBook Pro in drive tests, when it came to tests that took advantage of HyperThreading, the Core i5 MacBook Pro was twice as fast as the 13" MBA and three time as fast as the 11" MBA.
In their review of the 11" MacBook Air, PC Mag benchmarked the MBA against its ultraportable PC counterparts and found that while it was impressive for 3D gaming, it fell behind in several other categories.
The reason why Apple chose to stay with a Core 2 Duo was so that it can benefit from a better graphics environment, specifically Nvidia's integrated one. The Nvidia GeForce 320M graphics chip is the MacBook Air's one redeeming feature in terms of performance. Though it's not the kind of laptop you'd bring to a LAN party, it's a better gaming solution than Intel's integrated graphics-the kind found in the Asus Ul20FT-A1, Acer AS1830T-3721, and Toshiba T235-S1350. Its 3DMark 06 scores (4,569 and 3,984) were at least three-times better than the rest of the field. It was the only laptop that could handle our 3D intensive gaming demos, Crysis and Lost Planet 2.
But overall, they concluded:
the [11"] MacBook Air is not the zippiest laptop. It took almost four times (23 minutes 23 seconds) as long to encode a video than the Toshiba T235-S1350 (6:24). Because the Photoshop CS5 test is memory intensive, the MacBook Air (14:03) trailed against the Asus UL20FT-A1 (9:31) and Toshiba T235-S1350 (11:28).
The 13" MacBook Air can probably replace your main machine; the 11" MacBook probably not. The 13" MacBook Air can go toe-to-toe and in many cases outperform the current 13" MacBook Pro, but if you're going to be doing heavy video work and aren't afraid to go to 15", the Core i5 MacBook Pros still have the best performance. The 11" MBA is no slouch, but it's not nearly as convincing as a main-machine contender as the 13", especially for multimedia-heavy tasks. It's worth noting that the new MacBook Airs' increased resolutions make them seem a bit bigger than they are, in terms of what you can pack on the display—the 11" is basically equivalent to the 13 MBP and the 13" Air equivalent to the 15" MBP.