Apple Research Questions AI Reasoning Models Just Days Before WWDC

A newly published Apple Machine Learning Research study has challenged the prevailing narrative around AI "reasoning" large-language models like OpenAI's o1 and Claude's thinking variants, revealing fundamental limitations that suggest these systems aren't truly reasoning at all.

ml research apple
For the study, rather than using standard math benchmarks that are prone to data contamination, Apple researchers designed controllable puzzle environments including Tower of Hanoi and River Crossing. This allowed a precise analysis of both the final answers and the internal reasoning traces across varying complexity levels, according to the researchers.

The results are striking, to say the least. All tested reasoning models – including o3-mini, DeepSeek-R1, and Claude 3.7 Sonnet – experienced complete accuracy collapse beyond certain complexity thresholds, and dropped to zero success rates despite having adequate computational resources. Counterintuitively, the models actually reduce their thinking effort as problems become more complex, suggesting fundamental scaling limitations rather than resource constraints.

Perhaps most damning, even when researchers provided complete solution algorithms, the models still failed at the same complexity points. Researchers say this indicates the limitation isn't in problem-solving strategy, but in basic logical step execution.

Models also showed puzzling inconsistencies – succeeding on problems requiring 100+ moves while failing on simpler puzzles needing only 11 moves.

The research highlights three distinct performance regimes: standard models surprisingly outperform reasoning models at low complexity, reasoning models show advantages at medium complexity, and both approaches fail completely at high complexity. The researchers' analysis of reasoning traces showed inefficient "overthinking" patterns, where models found correct solutions early but wasted computational budget exploring incorrect alternatives.

The take-home of Apple's findings is that current "reasoning" models rely on sophisticated pattern matching rather than genuine reasoning capabilities. It suggests that LLMs don't scale reasoning like humans do, overthinking easy problems and thinking less for harder ones.

The timing of the publication is notable, having emerged just days before WWDC 2025, where Apple is expected to limit its focus on AI in favor of new software designs and features, according to Bloomberg.

Popular Stories

iPhone Air

Report: 'Virtually No Demand' for iPhone Air

Wednesday October 22, 2025 3:22 am PDT by
Apple is "drastically" cutting production of the iPhone Air and shifting focus toward the iPhone 17 and iPhone 17 Pro models, Nikkei Asia reports. The business publication claims to have learned of a major cut to iPhone Air production motivated by weaker-than-expected consumer interest, nearly to "end of production levels." Despite early reports of the iPhone Air selling out within hours of...
sam sung auction

Former Apple Employee Sam Sung Changed His Name to Avoid Attention

Wednesday October 22, 2025 4:44 pm PDT by
Back in 2012, an Apple retail employee named Sam Sung went viral because his name is similar to Samsung, one of Apple's main competitors. In a recent interview with Business Insider, he detailed that period in his life, how Apple responded, and he explained why he ultimately changed his name. Someone posted an image of Sung's Apple business card on Reddit in 2012, and it spread rapidly....
iOS 26

iOS 26.1 Coming Soon With These 8 New Features for Your iPhone

Wednesday October 22, 2025 6:15 am PDT by
The upcoming iOS 26.1 update includes a handful of new features and changes for iPhones, including a toggle for changing the appearance of the Liquid Glass design, "slide to stop" for alarms in the Clock app, and more. iOS 26.1 is currently in beta testing. The update will likely be released in the first half of November, and it is compatible with the iPhone 11 series and newer, but some...
cadillac lyric infotainment

GM to Remove CarPlay from All Future Vehicles, Including Gas Cars

Wednesday October 22, 2025 11:34 am PDT by
General Motors began phasing out support for CarPlay in its electric vehicles back in 2023, leading to complaints from iPhone users, but the company has no plans to back down. In fact, GM is going further and plans to remove CarPlay from all future gas vehicles, too. In an interview with The Verge, GM CEO Mary Barra said that the company opted to prioritize its platform for EVs, but the...
All Screen iPhone 2027 Feature 1

Apple's Plan to Launch Three New iPhone Designs Allegedly Revealed

Wednesday October 22, 2025 6:24 am PDT by
Apple plans to launch a new type of iPhone every year for the foreseeable future, according to an Asia-based source. The detailed information was shared by the account "yeux1122" in a blog post on the Korean platform Naver, citing domestic trend and component research companies. Corroborating other reports, Apple will apparently launch its first foldable iPhone in 2026, featuring a...
ios 26 1 liquid glass opaque

iOS 26.1 Beta 4 Lets Users Control Liquid Glass Transparency with New Toggle

Monday October 20, 2025 10:57 am PDT by
With the fourth betas of iOS 26.1, iPadOS 26.1, and macOS 26.1, Apple has introduced a new setting that's designed to allow users to customize the look of Liquid Glass. The toggle lets users select from a clear look for Liquid Glass, or a tinted look. Clear is the current Liquid Glass design, which is more transparent and shows the background underneath buttons, bars, and menus, while tinted ...
All Screen iPhone 2027 Feature 1

Report: Apple to Skip 'iPhone 19' Name for 'iPhone 20'

Thursday October 23, 2025 4:28 am PDT by
Apple's new iPhone lineup launched in the fall of 2027 will be called the "iPhone 20" models, rather than the "iPhone 19," according to research firm Omdia. Speaking at a conference in Seoul (via ETNews), Omdia Chief Researcher Heo Moo-yeol corroborated rumors that Apple plans to move the launch of its standard iPhone to the first half of the year and provided some additional clarity about...
maxresdefault

Apple's iPhone Air Experiment Fails as Supply Chain Cuts Production by 80%

Wednesday October 22, 2025 10:48 am PDT by
iPhone Air demand failed to meet Apple's expectations and the company's supply chain is scaling back shipments and production, reports Apple analyst Ming-Chi Kuo. Subscribe to the MacRumors YouTube channel for more videos. Suppliers are expected to reduce capacity by more than 80 percent between now and the first quarter of 2026, and some components with longer lead times will be discontinued ...
apple wallet drivers license feature iPhone 15 pro

iPhone Driver's License Feature in Apple Wallet App Launches in Another U.S. State

Thursday October 23, 2025 7:44 am PDT by
In select U.S. states, residents can add their driver's license or state ID to the Wallet app on the iPhone and Apple Watch, providing a convenient and contactless way to display proof of identity or age at select airports and businesses, and in select apps. Starting today, the feature is available to residents of West Virginia. To set it up, open the Wallet app and tap on the plus sign in...

Top Rated Comments

citysnaps Avatar
20 weeks ago
I don't find this surprising at all.
Score: 24 Votes (Like | Disagree)
trip1ex Avatar
20 weeks ago
Breaking news. The people who pretended otherwise always had something to sell.
Score: 22 Votes (Like | Disagree)
zorinlynx Avatar
20 weeks ago
LLM GenAI is pretty garbage technology. The less time it takes people to realize this, the better.

Yes, it does have some niche uses. But people are trying to push it as a solution to everything and even as far as replacing human beings, and it's just not capable of that. Not only that, but why do we want to replace human beings? Especially in the arts? I'd rather look at things made by people. It doesn't matter how visually stunning something is; art has no soul if there is no artist.
Score: 22 Votes (Like | Disagree)
turbineseaplane Avatar
20 weeks ago
“….and now here’s Ashley to talk about some new Genmoji!”
Score: 18 Votes (Like | Disagree)
Orange Bat Avatar
20 weeks ago
Of course. “AI” is just a marketing term at this point, and not any kind of actual intelligence. These AIs are really just glorified search engines that steal peoples’ hard work and regurgitate that work as if the data is it’s own. We’re just living in an “AI bubble” that will burst sooner rather than later.
Score: 16 Votes (Like | Disagree)
Salty Pirate Avatar
20 weeks ago
So AI is nothing more than clever programing?
Score: 15 Votes (Like | Disagree)