Apple A20 Pro Geekbench Scores Leak With Blistering Single-Core Speeds

The Apple A20 Pro has made its first public appearance on Geekbench, and the numbers explain exactly why Apple spent so much time talking up “Super” CPU cores at its September event. The listing, tied to a device identified as iPhone19,3 (widely believed to be the iPhone 18 Pro Max), shows a chip that doesn’t just edge past the competition. It lands a knockout blow in single-core performance while holding its own almost everywhere else.

According to the Geekbench 6.7 result, the Apple A20 Pro scored 4,725 points in single-core testing and 12,575 points in multi-core testing. Geekbench’s system information shows a six-core CPU split into a 2+4 configuration, with the two performance cores clocking in at a remarkable 4.93 GHz. The test device carried 12GB of RAM, which lines up with earlier supply chain reporting on the iPhone 18 Pro Max. That single-core number is so far ahead of anything Apple has shipped before that it strongly suggests the “Super” cores inside the Apple A20 Pro borrow directly from the architecture Apple built for the Apple M6, even though no Geekbench listing for that chip exists yet to compare against directly.

Apple A20 Pro vs. Snapdragon 8 Elite Extreme Gen 6

Here’s how the Apple A20 Pro’s early Geekbench 6.7 numbers line up against Qualcomm’s most recent leaked scores:

The gap in single-core performance is where this leak gets genuinely interesting. Based on these individual runs, the Apple A20 Pro comes in around 38 percent faster in single-core performance than the Snapdragon 8 Elite Extreme Gen 6, despite that Qualcomm chip hitting an even higher peak clock speed of 5.01 GHz on paper. The lead holds up in multi-core testing too, where the Apple A20 Pro is roughly 35 percent faster despite running two fewer cores. Apple’s decision to stick with a smaller six-core CPU while Qualcomm leans on eight cores clearly hasn’t hurt it in tasks that depend on a single fast core, which covers a huge share of everyday smartphone use like app launches, web browsing, and UI responsiveness.

The GPU Gains Are Just as Significant

A separate Geekbench 6.7 Metal benchmark from the same test device scored 64,069 points and explicitly identified the graphics processor as the Apple A20 Pro GPU. Here’s how that compares to last year’s chip:

That works out to roughly a 40 percent generational improvement in this particular test, and it matches almost exactly what Apple claimed on stage during its official unveiling of the Apple A20 Pro. The company promised up to 40 percent faster graphics performance thanks to the new seven-core GPU design, one more core than the six-core GPU found in the Apple A19 Pro. Apple also confirmed the chip is built on a 2 nanometer process, widely believed to be TSMC’s N2 node, along with 32 Neural Engine cores (double the previous generation) and 50 percent more memory bandwidth.

Update: Apple A20 Pro Breaks 4,000 Points on Geekbench 7

A fresh Geekbench 7 listing has since surfaced for a device identified as iPhone19,2, a different model number from the iPhone19,3 unit above and likely the standard Apple iPhone 18 Pro rather than the Pro Max. This result puts the Apple A20 Pro at 4,006 points in single-core testing and 11,460 points in multi-core testing, and it currently sits at the top of Geekbench’s own public single-core leaderboard for any chip, phone, or computer. No other processor in that database, including Apple’s own Apple M5 Max, has broken the 4,000-point barrier in single-core testing on Geekbench 7.

Here’s how that stacks up against the Apple A19 Pro, the Apple M5 Max, and AMD’s dual 3D V-Cache flagship, the AMD Ryzen 9 9950X3D2:

That’s roughly a 25 percent single-core gain over the Apple A19 Pro on this newer benchmark version, and it means the Apple A20 Pro is now outscoring desktop-class silicon like the AMD Ryzen 9 9950X3D2, a 16-core, 32-thread chip built for high-end gaming PCs. It’s a genuinely rare moment where a phone chip tops a fresh benchmark chart ahead of everything else Geekbench tracks, including Apple’s own Mac silicon.

Update: The 32-Core Neural Engine Shows Its Strength in Geekbench AI

Apple doubled the Neural Engine core count on the Apple A20 Pro, going from 16 cores on the Apple A19 Pro to 32 cores, and a Geekbench AI 1.7.0 result for the same iPhone19,2 device shows exactly why that mattered. The listing puts the chip at 7,002 points in the Single Precision test, 55,720 in Half Precision, and 75,810 in Quantized testing, all of which measure how quickly the Neural Engine handles machine learning workloads like image recognition and on-device AI processing.

Here’s how those scores compare to the last few generations of Apple silicon, along with the Apple M5 chip found in Apple’s latest Mac lineup:

The Single Precision gain over the Apple A19 Pro works out to roughly 27 percent, while the Quantized score jumps by close to 45 percent, and both numbers put the Apple A20 Pro ahead of the Apple M5 chip despite that chip powering full-sized Macs rather than a phone. That’s a meaningful result because Quantized performance in particular reflects the kind of 8-bit inference work that on-device AI features rely on, from Apple Intelligence’s local processing to third-party apps doing real-time image and language tasks. Doubling the Neural Engine core count clearly wasn’t just a spec sheet bullet point this year.

It’s worth remembering that Geekbench results captured this early can vary from unit to unit and don’t always reflect sustained, real-world performance once thermal limits kick in during longer sessions. Apple’s own claims about the Apple A20 Pro delivering 40 percent more sustained performance than the Apple A19 Pro will need independent, hands-on testing to confirm once review units are widely available. For now, though, this first synthetic benchmark run gives Android chipmakers a very clear target to chase, and it’s one that keeps moving further out of reach.

Stay connected via Google News

Follow us for the latest travel updates and guides.