New version of Gemini beats other AIs at math, science, and reasoning
Google's new Gemini Pro is smarter than other AIs at reasoning, science, and coding.
This is according to a series of benchmark results posted by Google on Thursday. In short, Gemini 2.5 Pro beats chief competitors at nearly everything — though we're sure the companies behind those competitors would disagree.
According to Google's data, Gemini 2.5 Pro has a healthy lead over OpenAI o3, Claude Opus 4, Grok 3 Beta, and DeepSeek R1, in the Humanity's Last Exam benchmark, which evaluates a model's math, science, knowledge, and reasoning. It's also better at code editing (per the Aider Polyglot benchmark), and it wins over all competitors in several factuality benchmarks including FACTS Grounding, meaning it's less likely to provide factually inaccurate text.
The only benchmark in which Gemini 2.5 Pro isn't a clear winner is the mathematics-focused AIME 2025, and even there the differences between results are pretty small.
SEE ALSO: Gemini now autogenerates summaries for long Gmail threads
As a result of all the improvements in Gemini 2.5 Pro, this model is now on top of the LMArena leaderboard with a score of 1470.
There's a catch, though: The final version of Gemini 2.5 Pro isn't widely available yet. Google calls this latest version an "upgraded preview," with a stable version coming "in a couple of weeks." The preview should now be available in the Gemini app, though.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Digital Trends
an hour ago
- Digital Trends
Apple's new iPadOS 26 shows Google what Android tablets need
Since the launch of the first iPad, Android tablets have struggled to compete. A lack of optimized apps means that Android tablets still don't offer as seamless a transition to larger displays, which also impacts the best folding phones, where many apps don't take advantage of the larger displays. However, one area where Android tablets have outperformed the iPad is in productivity and multitasking. Samsung's DeX desktop mode has been key to ensuring that some Android tablets offer a desktop-like experience, albeit one that needs more optimized applications. Last week, OnePlus also joined the party with the innovative Open Canvas system on the OnePlus Pad 3. Recommended Videos Until WWDC 2025, Apple had basic multitasking features like SlideOver, despite the hardware powering its best iPads being the same chipset as the new MacBook Air. Last year's iPad Pro M4 (2024) is extremely powerful and ideally suited to the new iPadOS 26 experience. Two days into using it, I've already covered how it's transformed the iPad. However, I've also noticed three things that Google should implement immediately on all Android tablets, as well as a broader problem it needs to address. The new windowing system and optimized apps To varying degrees, the iPad and all Android tablets share a common problem: they both employ a mobile-first platform, adapted for larger screens. As a result, neither quite achieved the same desktop experience offered by the best Windows tablets, which utilize a desktop-class platform optimized for smaller tablet sizes. The new windowing system changes this, and brings the iPad much closer to the Mac. Interestingly, Google also just announced that a new windowing system is coming to Android tablets in Android 16. Built upon the Samsung DeX experience, it's designed to revamp the Android tablet experience, but strangely, it won't be available on folding phones. Instead, you'll need to plug one of the best Android phones into an external display to activate the same feature. Each manufacturer can choose whether to implement Android 16's desktop mode or use an alternative, if at all, so we can still expect to see some differences in multitasking on Android tablets. Having tried the current desktop-mode implementations on Android, it's clear that alongside this feature, Google needs to entice developers to build apps and experiences that can properly transition from a touchscreen tablet to a desktop mode when docked or paired with a keyboard accessory. The menu bar and cursor I've been surprised at how crucial the new menu bar and cursor are to the new desktop-like experience. Two days in, I'm surprised at how quickly it felt familiar, how it feels more desktop-like, and how I already default to looking at the menu bar for options. The experience makes the iPad more akin to the Microsoft Surface than it does to the less-optimized experience offered by Android tablets. The menu bar and revamped cursor also make the iPad feel more like a small Mac rather than a big iPhone. This is a key distinction, as it comes in direct response to criticisms and feedback from iPad owners and the media, who are asking for better apps and features to take advantage of the increased power in the iPad. The new platform goes a long way toward addressing this, especially since the new Files and Preview apps are optimized versions of their Mac counterparts, bringing all the best features to the iPad for the first time. Apple has just announced iPadOS 26, so we won't see many apps build new experiences for it yet, but already, every app has a menu bar. It's filled with all the familiar commands from the same apps on Mac, and instantly familiar even if you use a Windows PC. A seamless transition between desktop and tablet Apple has managed to achieve what every tablet maker has wanted to: an experience that feels seamless whether using it with a keyboard or not. It's powerful enough with a keyboard attached to replace your Mac for many tasks, yet it's designed to work just as well with a finger. I've used the Galaxy Tab S10 Ultra and the OnePlus Pad 3, and I enjoy using both tablets for different reasons. The former is one of the best tablets that Samsung has ever produced, featuring DeX, a best-in-class solution on Android. The latter is an incredibly polished tablet with great specs, offers outstanding value for money, and has the Open Canvas multitasking solution. Each of these is excellent when used as an Android tablet, but neither matches the productivity of a PC or Mac in desktop mode. After installing iPadOS 26 on my iPad Pro, I have instantly become more productive, and it has become my favorite small laptop to use. An alternative to the Apple Magic Keyboard Even if Google were to adopt a similar system and approach to iPadOS 26, there's one other area that Android tablets regularly fall short: accessories. In particular, the Magic Keyboard for the iPad. Available for most of Apple's iPad range, the Magic Keyboard transforms the iPad into a mini laptop by using magnets to suspend it above the keyboard and bring the display closer to your eyes. It essentially mimics the display and keyboard setup of a computer, making it ideal for use on the go. Having used many Android tablets, there's one area that they seem to fall short: the keyboard. Most have adopted a Smart Keyboard or Folio case with an integrated keyboard, but these are ineffective when used on a lap, such as in a car or on a train. The Magic Keyboard holds the iPad in place even in these use cases, making it the ideal small computer for cramped spaces. Tablets have always required some form of tradeoff, especially as they're not as portable as a phone, and usually not quite as capable as a laptop. Microsoft addressed this issue with Windows, and now Apple has resolved it with iPadOS 16. Now, let's hope Google also does the same.
Engadget
an hour ago
- Engadget
Google will reduce battery life for some Pixel 6a phones to prevent overheating
Sam Rutherford for Engadget Google's Pixel 6a has been a winner in many respects, but a new issue with the smartphone's battery may pose a safety concern. After Android Authority reported on a pair of instances where a Pixel 6a battery overheated to the point where the device caught fire, the publication spotted some details in the latest Android 16 beta indicating that as a precaution, the Pixel 6a battery capacity and charging speed will be reduced. In addition, Android Authority received this statement from Google: A subset of Pixel 6a phones will require a mandatory software update to reduce the risk of potential battery overheating. The update will enable battery management features that will reduce capacity and charging performance after the battery reaches 400 charge cycles. We'll contact impacted customers next month, with all the information they need to address the issue. The Pixel 6a is due to receive software support until July 2027, but it's just the latest entry in this device line to have battery problems this year. In January, Google rolled out an update that reduced battery life of the Pixel 4a to increase stability for the 2020 phone model. Then in April, the company launched a program of repairs and payments after some instances of batteries "swelling" in the Pixel 7a.
The Verge
an hour ago
- The Verge
Posted Jun 11, 2025 at 4:51 PM EDT
Emma Roth Google has appointed a chief AI architect. Koray Kavukcuoglu, Google DeepMind's chief technology officer, will work to combine Google's AI models with its products 'with the goal of more seamless integration, faster iteration, and greater efficiency,' according to a memo seen by Semafor. Kavukcuoglu will reportedly remain Google DeepMind CTO while in his new role.
