In partnership with
🖱
Chinese AI models just crushed their American rivals on OpenRouter processing 4x more tokens in a single week.
DeepSeek alone exploded 801% week-over-week, proving China's AI infrastructure isn't just catching up it's already ahead.
But wait, there's more:
Apple quietly dropped the M6 iMac with dual Neural Engines
OpenAI claims three major science breakthroughs and wants academics to "shepherd" their AI research
I’m Alex. Welcome to L8R by Innov8.
Let’s dive deep.
In today’s post:
Chinese AI models crush US rivals 4-to-1 on OpenRouter
Apple's M6 iMac arrives with dual Neural Engines
OpenAI claims three science breakthroughs, invites skeptics to check the math $699
Chinese AI Models Crush the US on OpenRouter
Chinese AI models are absolutely dominating the US right now.
Developers are routing four times as many tokens through Chinese models on OpenRouter.
DeepSeek is leading the charge with a massive 801% jump in usage.
The details:
DeepSeek's updated V4.1-Flash-1017 model processed a crazy 33.6 trillion tokens between October 19 and 25.
Chinese models grabbed 49.8% of all OpenRouter traffic, while US models only held 11.9%.
The top five models on the platform are all Chinese, including tools from MiniMax, Zhipu AI, and Tencent.
MiniMax just dropped its M3.1 Flash model on October 19, making it three times faster and 70% cheaper.
This marks the 26th week in a row that Chinese AI labs have beaten US labs in token volume.
Why it matters:
Developers care about price and speed above all else.
US companies might make more total money, but builders are voting with their API keys.
They use Chinese open-weight models for heavy, high-volume tasks like coding and automation.
The gap in developer preference is only getting wider.
💡 L8R's Take:
The US needs to wake up right now.
You cannot win the global AI race if developers cannot afford to use your models at scale.
DeepSeek and MiniMax are proving that cheap, fast, and good is the ultimate winning combo.
From Our Partner
Your Search Bar Is Haunted
Shoppers type exactly what they want and get “0 results found” back. They don’t complain. They just vanish to a competitor.
Gladly understands what shoppers mean, typos and all, and turns dead-end searches into product recommendations.
Apple Drops the Wild New M6 iMac Built for AI
Apple just supercharged its 24-inch iMac with the brand new M6 chip.
This colorful desktop is now a total beast for running local AI models right on your desk.
The outside looks exactly the same, but the brains inside are completely new.
The details:
The new M6 chip packs up to a 12-core CPU and a 12-core GPU.
It features a massive Dual 16-core Neural Engine that makes on-device AI up to five times faster than the older M4 model.
Prices start at $1,499 for the base 10-core model and go up to $1,699 for the 12-core version.
Apple upgraded the connectivity with Wi-Fi 7 and Bluetooth 6 for blazing fast wireless speeds.
Pre-orders start on October 28, 2026, and the computers hit store shelves on November 6.
Why it matters:
Apple is getting very serious about AI.
The dual Neural Engine and new GPU accelerators mean you do not need the cloud to run heavy Large Language Models.
They want the iMac to be the ultimate local AI machine for everyday users.
This is a massive leap for anyone still holding onto an old Intel Mac.
💡 L8R's Take:
I love the raw power of the M6 chip, but keeping the exact same design is a little lazy.
Still, getting this much on-device AI muscle for $1,499 is a steal.
If you skipped the M3 or M4, this is exactly the desktop upgrade you have been waiting for.
From Our Partners
The AI notetaker that gets the hard words right
Your AI is only as good as what you feed it. Feed it a transcript with your product names, acronyms, and numbers spelled wrong, and you spend the afternoon hand-fixing the output.
Wispr Flow Notetaker uses your dictionary and calendar, so product names, acronyms, numbers, and uncommon names come out spelled right. It pops up when your meeting starts and captures Zoom, Google Meet, Teams, and Slack huddles with one click, no bot joining the call.
Then your meetings show up inside Claude or ChatGPT through the built-in connector. No copy paste. Try Notetaker free on Mac and Windows.
OpenAI Claims Three Massive Science Breakthroughs
OpenAI just dropped a bombshell on the science world.
Their internal AI models claim to have solved three massive problems in math, computer science, and physics.
To prove they aren't bluffing, they are handing the raw data over to top human scientists to verify.
The details:
The AI proved the 1991 local smoothing conjecture, which also solves two other massive math mysteries.
The models claim to have disproved the Exponential Time Hypothesis, a core rule that hundreds of computer science papers rely on.
The AI generated a proof for pure Yang-Mills theory, which is one of the famous Millennium Prize problems in physics.
OpenAI is using a new shepherd model where top experts, like Nobel laureate David Gross, verify and explain the AI's work.
OpenAI announced on October 26 that it does not want academic credit for these discoveries and will not make scientists sign NDAs.
Why it matters:
OpenAI got a lot of hate in the past for dumping unverified math papers online.
Now, they are playing nice with the academic world.
If these AI proofs turn out to be true, it changes everything.
We could see much faster computer algorithms and a totally new understanding of quantum physics.
💡 L8R's Take:
This is exactly how AI companies should handle big discoveries.
Let the AI do the heavy lifting, but let human experts double-check the math before calling it a win.
If these proofs hold up, we just entered a new era of AI-driven science.
🚀 Quick L8R Summary
Chinese AI surge: DeepSeek exploded 801% as Chinese models crushed US ones 4-to-1 on OpenRouter this week.
M6 iMac: Apple dropped a new 24-inch iMac with the M6 chip dual Neural Engines and smarter GPU cores for $1,499.
OpenAI science claims: OpenAI says its models cracked three big problems in math, CS, and physics now bringing in academics to verify.
📩 Innathe L8R engane undarunnu 👇?
We read every reply - just reply to this email and let us know how we can improve !
Appo adutha L8R il kanaam bie…👋