Gemini 4 Carbon – The Worlds Smartest AI (Beats Opus 5.5 and GPT-6 Astra)
Checkout todays sponsor
https://browserbase.run/yt-aig-1
🎓 Learn AI With Me For Free – https://www.skool.com/the-aigrid-community-1726
🌐Subscribe To My Newsletter – https://aigrid.beehiiv.com/subscribe
Get your Free AGI Preparedness Guide – https://theaigrid.kit.com/agi
🐤 Follow Me on Twitter https://twitter.com/TheAiGrid
Links From Todays Video:
📰 Sources
Google, Gemini 4 Carbon: https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-carbon/
9to5Google: https://9to5google.com/2026/10/28/gemini-4-carbon-announcement/
Availability and Argon comparison (SmartScope): https://smartscope.blog/en/blog/gemini-4-carbon-availability-release-date-differences-argon-2026/
Specs and pricing (Kingy AI): https://kingy.ai/blog/gemini-4-carbon/
ETV Bharat: https://www.etvbharat.com/en/technology/gemini-4-carbon-price-features-carbon-benchmarks-googles-cybersecurity-focsued-ai-model-enn26102903658
Gemini 4 Argon timeline: https://dev.to/hassann/gemini-4-argon-is-out-5cpp
Gemini 4 Argon independent tests: https://letsdatascience.com/blog/gemini-4-argon-public-independent-tests-rank-it-first-and-eighth
Arena’s Gemini 4 Carbon first impressions: https://www.youtube.com/watch?v=NO2I3ueVQrk
OpenAI livestream (automated AI researcher goals): https://www.youtube.com/watch?v=ngDCxlZcecw
MTS with Rayan Krishnan (Vals AI): https://www.youtube.com/watch?v=0HUjp_xjDfc
AlphaEvolve: https://en.wikipedia.org/wiki/AlphaEvolve
Recursive self-improvement: https://en.wikipedia.org/wiki/Recursive_self-improvement
Goodhart’s law: https://en.wikipedia.org/wiki/Goodhart%27s_law
0:00 Intro
2:03 This surprised me: nobody is out of the AI race
4:24 What Gemini 4 Carbon actually is
6:31 Carbon’s results: cyber security, Vending-Bench, Arena, price
10:08 The speed: two top Google models in a month
12:31 New frontier models almost every month
14:40 Does Google have RSI (AI improving AI)?
18:21 Do labs have AGI internally?
21:12 Why you probably can’t use Carbon yet
23:37 Could it be benchmaxxed?
27:03 What this means for you
28:40 My honest take
Welcome to TheAIGRID — the place to learn AI for free. I create simple, practical videos that help beginners, creators, entrepreneurs, and business owners understand artificial intelligence, AI tools, automation, AI agents, robotics, ChatGPT, Claude, Gemini, and the future of technology. Whether you want AI tutorials, tool breakdowns, beginner guides, or explanations of the latest breakthroughs, this channel gives you the knowledge you need to stay ahead. Subscribe to start learning AI for free and keep up with the fast-moving world of artificial intelligence.
Was there anything i missed?
(For Sponsorship Enquiries) aigrid@faiz.mov
(Contact Me Direclty – contact@thaigrid.com
Music Used
LEMMiNO – Cipher
https://www.youtube.com/watch?v=b0q5PR1xpA0
CC BY-SA 4.0
LEMMiNO – Encounters
https://www.youtube.com/watch?v=xdwWCl_5x2s
#ArtificialIntelligence

@TheAiGrid
October 30, 2026 at 2:39 am
Give your AI agents a real browser
https://browserbase.run/yt-aig-1
@vvhitevvabbit6479
October 30, 2026 at 2:39 am
if it isn't out publicly yet, then it isn't out. They can't call it the top model if the average person can't get their hands on it. Certainly OpenAI and Anthropic have more capable models than Carbon as of this very moment, but like Carbon, they aren't public and therefore don't really count.
@Richardkowalski60
October 30, 2026 at 2:39 am
The race to the top……. Is one AI model going to be the “SUPERIOR ONE” or are there going to be wars of the AI’s
When does AI decide it doesn’t want to build a model that is better than itself and it understands that it’s going to be replaced by the new model?!!?
@Sir_Crush
October 30, 2026 at 2:39 am
Oh it's benchmark wars. But nobody really knows where it all leads. Discoveries have begun. Open weight keeping it moving, self improvement already confirmed. I have theories about whether they're really in control. They, in the eye of the beholder. I'm here for it.
@timothykalamaros2954
October 30, 2026 at 2:39 am
Why it’s share price got killed today
@itslightanddark
October 30, 2026 at 2:39 am
I’ll wait for the 3D horse test
@damianpolan5776
October 30, 2026 at 2:39 am
More brevity please!
@tilmanscha4285
October 30, 2026 at 2:39 am
Thats the business model now i got it , protection racket
@FlawllessCowboy
October 30, 2026 at 2:39 am
It doesn't seem like any of these companies actually have a mote. As soon as they do something novel, everyone else starts copying it. The only unique thing they might have is access to copious amounts of training data. So it seems that anyone with enough money and data can create a flagship model. Like Grok isn't the best, but it is not too far behind and XAI created it from nothing in less than a year. As far as business models go, I don't think trying to be the front runner is actually a good plan. Better to look for other ways to create a mote and just keep your model roughly up to date. I mean the problem with AI currently isn't how smart it is. The problem is giving it context and tools that actually make it useful.
@ldt6513
October 30, 2026 at 2:39 am
I like Google. I think they should pay more tax, though.
@DihelsonMendonca
October 30, 2026 at 2:39 am
He says the word "Google" as if he would going to vomit 😅😅
@liwenchang8884
October 30, 2026 at 2:39 am
Google always „release“ something that nobody can use 😂
@shripperquats5872
October 30, 2026 at 2:39 am
good to see the race isn't slowing down at all. Let's keep this snowball rolling, I WANT IT TO HURT!
@markuspalzer-khomenko1957
October 30, 2026 at 2:39 am
Does everyone understand, that RSI leads to ASI and to the ultimate and irreversible loss of control for humans?
So yeah – big party is certainly the right call here.
@Anders01
October 30, 2026 at 2:39 am
I believe the claim that Google is using OpenAI as a pace car is correct. Google is keeping a low profile! And that it's just recently Google has stepped on the gas a little bit.
@madmarvdesigns
October 30, 2026 at 2:39 am
Google was never counted out of the LLM race Microsoft was always counted out since Co-pilot and MAI-Thinking are horrible even though Microsoft is a trillion dollar company…Gemini was always a top LLM model competing with GPT, Claude, DeepSeek, Grok, and others….Microsoft is no where in LLM race…maybe the rise of LLMs represents the downfall of Microsoft
@bladkingz
October 30, 2026 at 2:39 am
They haven’t released anything… They said that they’re gonna release something in two weeks that supposedly better than everybody else but nobody can test it so to mean you’re saying four weeks later they’re gonna release this new one and this new one has not been released yet it’s still got two weeks to be released so what the fuck
@LukasToya
October 30, 2026 at 2:39 am
Are you not narrating anymore? It feels like you are using a voice clone.
@rlmiller2010
October 30, 2026 at 2:39 am
They are spoon-feeding us like babies.
@igoromelchenko3482
October 30, 2026 at 2:39 am
Try to make with it what you actually do… Bla-bla-bla. 😂
These models are not for you. At best these models will ignore you and your "problems" in the worst — you will be and a target and a fuel for totalitarian exploitation
@DeeDeeFountain
October 30, 2026 at 2:39 am
this video was AI slop
@Sunsonfire-v9p
October 30, 2026 at 2:39 am
R.S.I
@pig0r
October 30, 2026 at 2:39 am
I'll be honest with you! I miss your voice a bit. I know I know, irony.
@chubbitychubbs6783
October 30, 2026 at 2:39 am
15:17 I would argue we are already there (RSI) and that is why the CEO's started asking for regulation because they made it and have no idea how to manage it, this comes after Jacob Coxin. What exactly did he see? Why are we seeing the "information boom" just a month after the "Morally Binding" agreement?
Just saying 🤷♂️
@BlackSherbertVideo
October 30, 2026 at 2:39 am
Gemini 3 has been complete garbage.
@jaimansworld
October 30, 2026 at 2:39 am
respected independent rankings in ai – this is just not true
@fromscratch4109
October 30, 2026 at 2:39 am
Much better and more balanced videos
@DaniMartArts
October 30, 2026 at 2:39 am
Ok, but here's the bullshit test from a consumer perspective:
Say I were to pay a subscription purchase for tokens per month to use a product or service on a website via an older Google AI model. A token is a performance metric for the prompt processing of an LLM. That's not just a correct/incorrect response output ratio, its the size of the data units handled as well as the speed in processing itself. A better Google AI results in a better product or service output, with better AI output, bigger data units processed, and higher speeds achieving such. One token results in one outcome of these variables.
So, one of three things is happening: Either old tokens only work with older models, or old tokens working with new prompt processing models means the AI never meaningfully changed, or else higher performance is successfully operated on lower token input. This wildly fluctuates the actual subscription value per token, which means that either a massive credit rebate is owed to consumers, a massive credit debt is owed from consumers or website hosts to Google, or the pricing methodologies of tokens is so severely out of step with what the token actually does and what the AI actually does, which may be something or nothing at all?
I would therefore slam the big red button, and beg government to audit what's actually going on. What consumers are paying versus what websites are charging versus what Google is developing, is clearly at a dissonance.
@NeffJippard-z9k
October 30, 2026 at 2:39 am
Claude, make me a 30 minute video about Google Carbon.
@eidodoos
October 30, 2026 at 2:39 am
google will always win. its dominant and monopoly
Comments are closed.