Resources:
Join the discussion: Got something to say? Let us know on LinkedIn and network with other AI leaders
Upcoming Episodes: Check out the upcoming Everyday AI Livestream lineup
Connect with Jordan Wilson: LinkedIn Profile
OpenAI’s “Code Red”: What Google Gemini’s Surge Means for AI Strategy in Business
The latest episode of the Everyday AI podcast dissected recent reports pointing to a “code red” moment inside OpenAI. Major benchmark wins and surging user growth for Google’s Gemini 3 have pushed OpenAI to reconsider its priorities, signaling real-time market dynamics that every business leader should analyze. For those weighing investments in AI platforms, understanding these developments can clarify decisions about product adoption, budget allocation, and future-proofing digital strategy.
AI Market Share Trends: Google Gemini Rapidly Closes in on ChatGPT
OpenAI once dominated the generative AI conversational market, controlling 87% of market share a year ago. Recent similarweb data shows that figure dropped to 71%, while Google Gemini climbed from 5.7% to 15%. This shift was largely attributed to November’s rapid-fire major AI product launches: OpenAI’s GPT-5.1Pro (limited to $200/month premium customers), Google’s Gemini 3 Pro (hailed for ‘shattering benchmarks’ like a 37.5 on humanity's last exam), and Anthropic’s Claude Opus 4.5.
The specific market share erosion for OpenAI wasn’t incremental—it arrived almost exclusively in the wake of Gemini’s launch. In business terms, platform traction is where contracts, internal adoption, and developer investment flow toward, making this data highly actionable for decision-makers considering AI integrations.
Benchmark Performance: AI Model Selection Matters
Benchmarking is now the leading metric used by enterprise buyers for AI adoption. Gemini 3 Pro overtook GPT-5.1 across all but one major benchmark, validating Google’s candid claim of “most intelligent model ever.” Claude Opus 4.5, meanwhile, set a new industry record by surpassing 80% on the Sweebench Verify coding benchmark—essential for businesses optimizing workflows and automating complex software engineering tasks.
Crucially, OpenAI’s models have lost top spots in comparative, blind taste tests (LMSys’ LM Arena). GPT-5.1 High—their premium “thinking model”—ranks sixth, with Gemini and Claude outperforming in critical business contexts such as coding and reasoning. The default ChatGPT model, used by most customers, slipped out of the top 20 in many analyses. For any company relying on out-of-the-box AI, this differentiation could result in material impacts on output quality, productivity, and risk management.
Default Model Quality: Impact on User Experience and Productivity
One of the most overlooked operational details from the show: over 93% of ChatGPT paid users continue to use the default model, which benchmark studies reveal to be markedly less capable than alternatives. Only 24% of paid users utilize the more advanced “thinking models”—despite the business-critical performance delta.
This has concrete implications: A business deploying workflows based on the default ChatGPT experience is not leveraging the full potential of LLM technology, possibly facing suboptimal outcomes in customer support, market research, or internal automation. Segmenting users to help them select and consistently apply the right model could directly affect competitive positioning.
Product Diversification: OpenAI’s Split Focus vs. Google’s Ecosystem Play
OpenAI’s breakneck expansion now covers five distinct avenues: frontier LLM development, consumer ecosystem (serving 800 million weekly active users), enterprise developer platforms, custom hardware ambitions, and trillion-dollar infrastructure commitments. This wide portfolio has led to projected financial strain—$22 billion in 2025 spending versus $13-$20 billion in anticipated revenue.
Compared to Google, which capitalizes on Gemini’s distribution across Cloud, Search, Android, Maps, and more, OpenAI’s risk comes from a less integrated ecosystem and split attention. For a business choosing a partner for AI-driven transformation, a vendor’s focus and infrastructure commitments are increasingly vital considerations.
Financial Health and Stickiness: Enterprise Adoption and Profit Pathways
Anthropic positions itself for profitability as early as 2028, while OpenAI may only break even by 2030, according to recent reporting. Notably, OpenAI’s “stickiness”—measured at 33% daily reliance—is unmatched, far outpacing competitors like Gemini (5%), Claude (16%), Copilot/Grok (14%). High engagement rates validate the advertising potential and strengthen the case for enterprise adoption when continuity and deep user context matter.
Supporting this, ChatGPT’s average query length is 60 words, compared to Google’s AI mode at 10.4 and classic search at 3.4. This richer context opens avenues for more targeted platform experiences, deeper insight extraction, and eventually, higher-value B2B and B2C commercial opportunities.
Actionable Insights for AI Investment and Adoption
Model Selection Is Critical: The majority of users—and possibly employees—may be using default models that underperform. Empowering teams to leverage advanced versions can tangibly affect results.
Ecosystem Integration Confers Resilience: Platforms like Gemini, embedded throughout enterprise tools and cloud infrastructure, offer distribution advantages that may tip the choice for future projects.
Stickiness Predicts Business Value: High engagement platforms like ChatGPT remain central to user habits, which translates to advertising, data richness, and retention leverage.
Track Financial Stability: OpenAI’s broad ambitions bring financial risk, but their dominant mind share and market stickiness indicate enduring value; Anthropic’s profitability focus may appeal to risk-sensitive B2B buyers.
Benchmark Data Should Guide Decisions: Don’t rely solely on market perception—track up-to-date comparative benchmarks to inform integrating AI for coding, decision support, or multi-modal applications.
The underlying lesson: The landscape is shifting rapidly, with Google Gemini mounting a credible challenge to OpenAI’s crown in key technical and adoption metrics. Strategic decisions around model selection, platform partner, and internal deployment should lean heavily on objective benchmarks and real engagement metrics rather than headline narratives.
For full market analysis and tactical advice, consult independent benchmarking platforms and demand clear answers from any vendor about which model powers your workflow—and whether your team is using it to its full potential.
Topics Covered in This Episode:
- OpenAI "Code Red" Panic Explained
- Google Gemini 3 vs ChatGPT Benchmarks
- November AI Model Release Impact
- Anthropic Claude Opus 4.5 Coding Leader
- OpenAI Financial Strain & Revenue Analysis
- Market Share Shift: ChatGPT vs Gemini
- ChatGPT Default Model Quality Critique
- Enterprise AI Adoption: OpenAI vs Anthropic
- OpenAI Ads Controversy and User Backlash
- App Stickiness and Query Length Data
Episode Transcript
Jordan Wilson [00:00:46]:
So on today's show, we'll answer if OpenAI is still the king of the AI hill or if this code red situation going on right now means that they've maybe lost the crown to Google.
Jordan Wilson [00:02:07]:
I'm excited to get into it. I hope you are too. Welcome to Everyday AI. What's going on y'all? My name is Jordan Wilson, and this is for you. This is your daily livestream podcast and free daily newsletter helping everyday business leaders like you and me not just keep up with all the AI news and developments. Yeah. You could tell by our opener there. There's a ton, but how we can make sense of them and make sound decisions to grow our companies and our careers.
Jordan Wilson [00:02:29]:
So if you're joining me on that journey, it starts here with the unedited, unscripted livestream podcast. But if you wanna take it to the next level, make sure you go to our website. That's your cheat code. It is youreverydayai.com. We're gonna be in our free daily newsletter, which you can sign up for there. We're gonna be recapping the highlights from today's show as well as bring you all of the other AI news that you need to know. Alright. Now let's get into it.
Jordan Wilson [00:02:52]:
Here's what you need to know about the new reports that say OpenAI is kind of in panic mode. They're hitting that siren. It is code red. They are losing ground. So on today's show, here's what we're gonna dive into. We're gonna investigate the internal Code Red panic inside of OpenAI following a brutal November. We're gonna break down the data showing how Google and even Anthropic have overtaken ChatGPT in some key areas and answer the ultimate question. Is ChatGPT's default era over?
Jordan Wilson [00:03:31]:
Is Google taking ChatGPT's crown? Well, kind of, but also no. Not at all. And here's the explanation in that. I think that over the past three years, right, we are now more than three years into the ChatGPT era, which has obviously come to define AI. And I think that right there is enough to say that, no, ChatGPT is not going to be losing its crown at least when it comes to mindshare, when it comes to just the sheer number of people that they have using and relying on the platform every day.
Jordan Wilson [00:04:17]:
we have agents. We have business, data that is synced dynamically, right into these front-end LLMs.
Jordan Wilson [00:04:53]:
So the space has changed so drastically over the past three years since ChatGPT kind of introduced us to the space, which is ironic since Google actually invented the transformer and had a kind of a technically a working version of something like ChatGPT public before ChatGPT even did. So, while I do think Google may permanently take the lead in LLM development, and I'd say in 2025, they've been the winner hands down.
Jordan Wilson [00:06:15]:
Google Gemini three overtook GPT 5.1 on all of them except one. So almost a clean sweep for Gemini. Gemini's new model, Gemini three pro.
Jordan Wilson [00:06:48]:
Right? So even before all of this, and I'm gonna get into, some of the discrepancies and how we use ChatGPT and what model you're selecting. Right. But if I'm being honest, the default model of ChatGPT, not really good. Alright? And so they just got hit with, big release after big release, and that has reportedly caused OpenAI to hit that big, panic button. I'm wondering if in these big companies, is there actually a button? Can someone tell me. Right? Friends at Big Tech. Let me know. That'd be great if there was.
Jordan Wilson [00:07:23]:
It'd be so cool. But I just I just imagine. Right? Like like, someone just sweating in the back and being like, yep. Gotta pop up the button. Time to hit it. But, you know, this is for the first time, I'd say, since, ChatGPT came out at the November 2022. I say it's the first time that their position has been questioned as the overall AI leader. I would say even, you know, mid November, early November of this year, so five weeks ago, I'd say if you've looked at everything across the board, you would say, nope.
Jordan Wilson [00:07:57]:
OpenAI is still comfortably in the lead. I don't think you could really say that anymore. And that's maybe what's caused OpenAI to get back to the basics here. So, here's what's important. Alright. So this is not the first code red because what's interestingly, what's interesting enough is Google was actually the first, company to kind of put out a code red when it came to AI. So a lot of people well, most people know Google invented the transformer, the underlying technology that ultimately led to, ChatGPT and other large language models, but people don't even know. Google had a working and live version of an AI chatbot, essentially, an in early preview version of Lambda, in their AI kitchen well before there was a ChatGPT, but it was extremely limited.
Jordan Wilson [00:08:58]:
They didn't put a lot of resources behind it. It was a research preview. Right? It was essentially for geeks, you know, kinda like me. So when this came, right, it was bad news for Google. I think Google thought that maybe, you know, this technology going mainstream was still maybe years away. Right. They were pioneers, in the technology, but the chat GBT moment. Right? And that's what it has become.
Jordan Wilson [00:09:28]:
And then we heard, maybe the December, and now all of a sudden, it might be December 9. So how the heck did this happen? Right? How did OpenAI go from, if I'm being honest, not really having any true competitors for the first twenty three and a half months and having an almost unfair advantage, right, when it came to market share, when it came to features, when it came to UI, UX, when it came to just branding. I mean, you talk to anyone, you say something about AI and, you know, everyone's like, you mean ChatGPT? Right? They, they, they beat everyone to the punch and had an insurmountable lead yet. Here we are being like, are they losing it? Have they lost it? And in two years, are they even going to be in the conversation of the one a one b? Well, here's one reason why I think we're in this position. OpenAI's focus for better or worse is much wider than it was a year ago. Right? You could make an argument depending on how you, categorize these things. You can make a an an argument now that OpenAI has five, you know, or OpenAI has five different businesses or has five very unrelated or not very much related, departments. So as an example, right, you have your Frontier large language models.
Jordan Wilson [00:10:47]:
So Google has grown more over the past twelve or thirteen months, and I don't even think that's close. So now it's the other side of the coin. Now it is OpenAI calling the, code red after Gemini three's launch and some other things. And like I said, this is the first time that I think public perception and even the markets, right, have said that maybe Google's the leader. Right? So Wells Fargo now prices Google stock as the AI race leader for the first time since 2016. So here's a little bit more on the details of this code red. So this, reportedly went into an internal memo, and it said, yo, we gotta stop with the bells, the whistles, the features, and we gotta get back to core model development. So this directive, cited temporary economic headwinds from Google's benchmark topping releases, and also might lead to a new OpenAI model getting released as early as Check Watch today.
Jordan Wilson [00:11:50]:
Right? So, report from The Verge said we might get a version of ChatGPT, maybe called GPT 5.2, as early as today. I've also, kind of heard or seen it could be Thursday. Right? We've seen the dates float around, but it could be as early as literally today. So if you are listening to this podcast, it's a live stream podcast, 07:30AM Central Standard Time in Chicago. Right? But maybe you're listening to this on your lunch break or your commute on the way home from work. There could already be a model out. Right? We'll obviously have that news if it when it does break in our newsletter. But this, essentially, all of this news accelerated, the launch of this new model because, we had heard previously that it might be released in January.
Jordan Wilson [00:12:36]:
And then we heard, maybe the December, and now all of a sudden, it might be December 9. So how the heck did this happen? Right? How did OpenAI go from, if I'm being honest, not really having any true competitors for the first twenty three and a half months and having an almost unfair advantage, right, when it came to market share, when it came to features, when it came to UI, UX, when it came to just branding. I mean, you talk to anyone, you say something about AI and, you know, everyone's like, you mean chat GPT? Right? They, they, they beat everyone to the punch and had an insurmountable lead yet. Here we are being like, are they losing it? Have they lost it? And in two years, are they even going to be in the conversation of the one a one b? Well, here's one reason why I think we're in this position. Opening eyes focus for better or worse is much wider than it was a year ago. Right? You could make an argument depending on how you, categorize these things. You can make a an an argument now that OpenAI has five, you know, or OpenAI has five different businesses or has five very unrelated or not very much related, departments. So as an example, right, you have your Frontier large language models.
Jordan Wilson [00:14:03]:
Right? That's your this is your GPT five series, the actual models being built. Right? Then you have your consumer ecosystem. Right? Because some of these players out there, they just make models, and that's it. Right? They put them out on the API or, whatever, but they're not building an ecosystem. Obviously, with 800,000,000 weekly active users, OpenAI is actively investing in a very vibrant ecosystem, and that comes with it a lot of challenges, a lot of those bells and whistles, a lot of these things that take time. Right? They have a browser, you know, their Atlas browser. They have the, you know, obviously, the ChatGPT app across, web, mobile, desktop, Sora with video, GPT image. Right? They have all these things.
Jordan Wilson [00:14:46]:
Then they have their enterprise and developer platforms. So the business and enterprise tiers of ChatGPT, the API side, the real time agents SDK, etcetera. Right? And then they have their new kind of in, upcoming hardware, division or ambitions. We'll see. Right? So custom AI chips. Right? they're getting in the chip game, like Nvidia, like Google with Google's, TPUs. Right? So it was reported that they're gonna be working with Broadcom, to develop an in house chip that they will use and probably be, selling or renting out to others. And then also on the hardware side, we've seen a lot of reports recently that we cover here on our Mondays new, Monday new shows.
Jordan Wilson [00:15:30]:
OpenAI has been pushing like everyone from Apple's hardware division, but it seems like, app, OpenAI is gonna be making a serious play in the AI hardware space. We don't know what the device is gonna be. We've seen reports. It's not gonna be a phone. It's not gonna be, you know, earbuds or glasses. It's gonna be something else, but they're building an AI hardware division, and they're getting real big players, to go with them. And then last but not least, they're an infrastructure company now. They are.
Jordan Wilson [00:15:58]:
They have more than a trillion dollars, of third party, in, infrastructure commitments. Right? Building, like, literally the Stargate, deal, with, Oracle and others. I mean, they're building just huge, huge data centers. The CapEx spend is enormous. Right? A $100,000,000,000, plus already committed with an NVIDIA deal slash investment people calling, you know, circular funding or vendor financing, whatever you wanna call it. Right? Those are areas that OpenAI was not even playing in for the most part a year ago. Right? Aside from, maybe the consumer ecosystem and a little bit on the, developer platform, but at least half of these things didn't really exist a year ago. So maybe this is the give and take that a, company must make in order to grow at the rate at which they want to grow.
Jordan Wilson [00:16:54]:
Sometimes you have to maybe throw, expensive spaghetti at the wall and see what's going to stick. But this split focus has led to reported financial strain. So, according to reports from the Wall Street Journal and Fortune, OpenAI projects a $22,000,000,000 spend versus a $13,000,000,000 in revenue in 2025. Although Sam Altman did say that that number in revenue is gonna be closer to 20,000,000,000 or a little above, yet it could anywhere be from, you know, couple billion dollars loss to up to $9,000,000,000 loss this year. And there have been estimates. Right? And this estimate was wild. This h, this HSBC, estimate that said, OpenAI may be meeting a $207,000,000,000 funding shortfall between now and 2030. Wow.
Jordan Wilson [00:17:47]:
So is this just a case of OpenAI being spread too thin? Is this an AI startup that, you know, maybe you should have just been an AI chatbot company and they wanted to take over the world and maybe they shouldn't have, and now they're losing their, you know, luster. They're, maybe, relinquishing what they were known for, or is it something more than that? Alright. We're gonna find out in a minute. Quick break. Word from our sponsors.
Jordan Wilson [00:19:24]:
app. So I would say part of this slipping from the top of the AI mountain for OpenAI is because of its split focus, but the timing of everything else has added to it undoubtedly. Right? So, November has been a pile up of new model releases. One one thing people don't talk about, well, OpenAI rolled out their own very powerful model in mid November GPT 5.1 pro. Right? But very few people have access to that because you have to be on the $200 a month plan. But then, obviously, the big news here came on November 18 when Google launched Gemini three Pro, calling it their most intelligent model ever. And then the next week, Anthropic followed up with Claude Opus 4.5, really finding its space, targeting coding agents and computer use. So let's quickly just look at Gemini three's reception, and it was, good.
Jordan Wilson [00:20:20]:
Right? People went nutty and, rightfully so, I'd say. When Gemini three came out, Gemini three Pro, it shattered every benchmark. Right? You know, getting scores on benchmarks that, you know, maybe a year ago, we thought would not be possible for multiple years. You know, one of them, scoring a 37.5 on humanities last exam. One of the newer, kind of, you know, datasets, or benchmarks since a lot of the previous ones like m m, MMMU, MMLU, a lot of these previous benchmarks have kinda been saturated. Right? Because all the models are, you know, scoring off the charts or, you know, close to a 100 on some of these. So it's like, alright. Some of them, some of them are saturated, but one that is not yet is humanity's last exam and Gemini just Gemini three leapfrogged in front of everyone.
Jordan Wilson [00:21:13]:
And then just their sheer user growth all the way up to 650,000,000 monthly active users, right, versus OpenAI's 800,000,000 weekly active users still. I think they've that's grown from, 400,000,000 just a couple of months ago. So Gemini is on fire, and Google is on fire in terms of user growth. And, pretty good anecdote we shared here, this quote, I think the day it came out or the day after it came out, Salesforce CEO, Mark Benioff, publicly switched and put out a tweet, that's like he said he's been using ChassisBD every year, spent a little bit of time with Gemini three, and he's like, I'm not going back. And I think that was a sentiment, kind of echoed by a lot of people. Right? It was that big of a leap, not just that big of a leap, but I think that, kind of the default chat mode inside of, ChatGPT has been stalled. Again, more on that here in a minute. So Google has significantly eaten into mark, OpenAI's market share.
Jordan Wilson [00:22:19]:
So this is, a graph here, for our live stream audience. You can see in podcast audience, if you ever wanted to know, I never have anything overly visual, but if you want the video version, I don't know. You can take screenshots of me making weird faces. You know, for the most part, you don't need a lot of video, but you can always see that on our website at youreverydayai.com. But you will see here, this is according to similar web data, which is extremely, accurate, looking at the market share over time, ChatGPT versus Google Gemini. So a year ago, OpenAI had 87%, of the market share. Now 71%. So a huge drop off.
Jordan Wilson [00:22:59]:
Whereas conversely, Google Gemini has been growing faster than anyone else. So a year ago, they only had 5.7% market share, and they've nearly tripled that now at 15%. So in the same period of time, you know, maybe, you know, some of the other players have maybe grown half percent, 1%. Right? Yeah. Google Gemini, tripled, and most of that, right, came from ChatGPT's lunch. So when I say that they're eating, ChatGPT's lunch on the model side and a little bit eating into their big pie, you know, the numbers support that. And also, the sleeping beauty herself, Anthropic finally woke up, from its summer slumber and decided to ship something, and buy something an extremely impressive model. Yeah.
Jordan Wilson [00:23:53]:
We gave it the live run through here on the show a couple of weeks ago. It failed some of our live tests, but that's why I always tell people. Right? You have to put it through your own, paces. Right? I think a lot of times when I'm doing, tests on models, it's sometimes when, the servers are overloaded, things get buggy. But Anthropic, Cloud Ovis 4.5, extremely impressive. Right? Especially, on any, software development, coding benchmarks. So the first model to ever get more than an 80% on the Swede Bench Verify, which is the industry's, kind of, you know, number one, real world real world coding benchmark. Right? And I'll probably do a show, at some point, maybe later this year or early next year.
Jordan Wilson [00:24:40]:
Just really talking about why I think coding, is more important, from a benchmark perspective and from a company perspective than most people realize. Right? And it's a little deep. Right? It gets into AGI and some other things, but I'll probably talk about that. So not only that, but anthropic finally. Right? And I said this on the show a year ago. I said anthropic is not going to survive as a company unless they reduce their prices to be competitive. Their API prices have never been competitive. And, obviously, if you're on a paid plan using cloud on the front end, you can't use it.
Jordan Wilson [00:25:14]:
It stinks. Right? The limits are so low, but they finally, cut their prices by 67%. So, again, when I said November was kind of a tough month for OpenAI, well, it was because its competitors came out and they came to play. They woke up for no reason and chose absolute violence. So speaking of Anthropic, the IPO rumors, did not help OpenAI's cause either. So this was last week just days before the reported, OpenAI code red. Some rumors on Anthropic, you know, laying the groundwork for an initial public offering. So this is from the Wall Street Journal.
Jordan Wilson [00:25:58]:
I'm gonna read an excerpt here. So it says, excuse me. Yeah. Still still fighting this little, this little bug, this little throat thing here. Got my medicine, my topo Chico. So it says Anthropic, which has a growing number of business users because of the capabilities of its quad chatbot encoding in other arenas, expects to break even for the first time in 2028, the documents show. By contrast, OpenAI forecast its operating losses that year to swell to about 74,000,000,000 or roughly three fourths of revenue, thanks to ballooning spend on computing costs. The ChatGPT maker also expects to burn through roughly 14 times as much cash as Anthropic before turning a profit in 2030.
Jordan Wilson [00:26:49]:
Alright. In layman's terms, here's what that means. And I've been saying this all along. Right? That I know that anthropic is much more profitable per customer, and I think that they're playing two very different games. Right? Anthropic is never going to catch OpenAI when it comes to users. Right? But what they are likely crushing OpenAI on is revenue per user. You know, even CEO, Sam Altman, for OpenAI has said that on many plans, they lose money by default. Right? Yet, this reporting is showing that anthropic is expecting to be profitable by 2028, which is pretty early because a lot of, earlier estimates said best case, OpenAI might not turn a profit until 2030.
Jordan Wilson [00:27:37]:
Right? Which is normal. Right? A lot of people look at this and, you you know, they they they put on their, you know, AI is a bubble. Oh my gosh. You know, Chad GPD is so popular, yet they're, you know, losing billions of dollars. That's how startups work. Right? I've I've given long, long receipts on this show, you know, companies that were, in the red for ten years or longer and are now the biggest and most profitable companies in the world. That's how it works. Right? If you're going to be a world changing company, you start as a start up.
Jordan Wilson [00:28:06]:
You take on debt, in for the most part, you're probably gonna be, burning through billions of dollars for five to fifteen years. That's how it goes. And hopefully, then you make it out on the other side as, you know, a fortune 100 fortune 50 company. And a lot of people don't realize by the path right now. OpenAI is the largest private company in the world. And if they were a public company today based on their evaluation, they would be a fortune 20 or a top 20 company in The US. Right? So they are no small beans and it is relatively normal to be burning through cash. However, when compared to anthropic, this might be another thing.
Jordan Wilson [00:28:45]:
Right? The Gemini three, released the Opus 4.5, and then the news, right, that OpenAI or sorry, Anthropic is seemingly being looking look looked at a little bit more favorably, on the markets and by, potential investors, than OpenAI is. So speaking of the public narrative punching OpenAI this week, there's another thing that happened in the last couple of days right before the code red. There's this kind of little we'll call it an ad controversy. Right? I don't know. I chuckle at this one, but it caused OpenAI a lot of bad press, and according to reports, ads is one avenue that they're going to pause on. So essentially, you had a lot of people complaining on Twitter. Right? And they're like, I'm a pay I'm a paid subscriber. Right? I'm paying $20.
Jordan Wilson [00:29:37]:
I'm paying $200 a month. And look at this. I'm getting an ad. It was a narrative that maybe OpenAI and I said so, myself, probably the UI UX, of this was not the best. It was actually an in app nudge. People might, argue, right, it's not technically an ad. Even if it might look like an ad and you it could be, construed as an ad. Right? A lot of this was, nudges usually unrelated to what you were searching for.
Jordan Wilson [00:30:06]:
Right? So usually an ad is something related to what you are searching for. Right? So if you're, you know, chatting with ChatGPT about the latest, I don't know, Bears game or doing some deep research on something and you get a little nudge for, installing the Peloton app on ChatGPT. Right? So it's not sending you to other websites or pushing products. It's trying to get you to, install ChatGPT apps. Right? A big play that they're trying to do there as well. So these are technically in-app nudges to try to get you to use ChatGPT apps. People lost their mind, and said, oh my gosh. I'm paying money, and here's these ads.
Jordan Wilson [00:30:45]:
Right? So opening, I did disable the feature, two days ago after chief research officer, Mark Chen, admitted on Twitter that quote unquote, we fell short and then they disabled the feature. So more or less, OpenAI didn't even put ads out there, and everyone lost their noodles. Right? And there's all these, you know, articles that came out in the press, about how this was just a bad look. So it's been a bad, tough tough couple of weeks, especially, the last maybe two weeks for OpenAI that has maybe led to this code red. But ready? Saving the hot take Tuesday, the hot take for near the end of the show here. I think here's the secret. Here's the key. Here's what no one's talking about.
Jordan Wilson [00:31:36]:
I think one of the reasons why ChatGPT is losing its ground is because the default chat model is bad. Right? And and, Sam Altman said in August that previously, only 1% of free users used a thinking model and only 7% of paid users used a thinking model. Those, numbers have increased, right, I believe to now up to 24% of paid users. But still, the overwhelming majority of everyone using ChatGPT is not using paid models or or sorry, not using thinking models. I would never.
Jordan Wilson [00:32:42]:
And my wife always gives me a hard time. Right? Because she's like, hey. Can you, you know, ChatGPT this quick? And I'm like, yeah. I can. Right. And then two minutes later, she's like, okay. What's the answer? And I'm like, you gotta give it time. Right? I mean, I I I gotta use the right model.
Jordan Wilson [00:32:57]:
Right? Otherwise, just go into, you you know, Google’s AI mode. Right? If if if you want the juice, you gotta wait. I would never I would never use the default, ChatGPT model for absolutely anything. And I, you know, might lose a couple of friends at OpenAI for saying that, but I don't think you should either because it's not good, and the stats show that. Alright. So let's look at LLM Arena.
Jordan Wilson [00:33:21]:
Talk about, unfortunately, OpenAI losing some of their ground, even their best model, their high thinking model right now is not a top five model. Alright. So it is at number six on LLM Arena. LLM arena is essentially a blind taste test for large language models. So it's not a top five model, which is crazy. It's crazy to say, because essentially all the way through the 2024, OpenAI was always the top model on LLM arena, and it was usually never even close. Alright. So now they've lost that.
Jordan Wilson [00:33:58]:
If you are using the default chat mode, right? The instant, or if sometimes if you're using auto and it doesn't just, it doesn't route you to a thinking model, you're not using a good model. You're using a poor model. Right? So ChatGPT five one is the fifteenth highest ranked model on LLM arena and the previous version ChatGPT five ranked twenty first.
Jordan Wilson [00:34:41]:
Right. It's not just LLM arena. If we look at artificial analysis, a great resource, third-party benchmarking, it, it's an aggregate of all these different benchmarks. Right? You'll see ChatGPT five one high, right there tied for second, with quad opus 4.5, but behind Gemini three Pro. Okay? So tied for second. ChatGPT five one high. Good model. Right? Default model, not even in top 20.
Jordan Wilson [00:35:14]:
Oof. That's one of the reasons. That's one of the reasons why Gemini and other people are taking the crown. Right? And that's why I think, right, my my Twitter feed is just literally just people talking about AI, and I've seen, story after story, people just being like, it seems like ChatGPT is not as good as it was before. And it's it's like, yes. Stop using the default model. Use the thinking mode. Right? That's why sometimes you gotta use multiple tools because, yeah, if I want something fast, I might just use Gemini three or use Google AI mode.
Jordan Wilson [00:35:49]:
Right? But sometimes if I want something, done the right way specifically, and if I wanna see in the chain of thought come on, Google. Please add this. In the chain of thought, if I wanna see where the model is deciding agentically to go into research, right, I can see that in ChatGPT five Pro or ChatGPT five thinking, right? I can't see that, unfortunately, in Gemini three Pro. So you shouldn't be using the default model ever until they improve it. Right? And maybe by this time that I'm done yapping, on this podcast, the ChatGPT five point two will be out or, you know, maybe later this week, and maybe it'll be a different story. Maybe it'll be a good base model. But that's one of the reasons why I think that so many people have just been like, ChatGPT is not very good anymore or why they're losing so much market share. And it hasn't really been talked about.
Jordan Wilson [00:36:42]:
Well, the default chat model that, reportedly, before 93% of even paid users were using. And now, about, 70, 76% of paid users, are still using that model. And it's not good. Right? I'm sorry. When I say it stinks, it's obviously world changing model. Right? You know, four years ago, if we ever would have thought we would have something like this, but comparatively, it's not good. Right? You don't wanna be, handing off important business tasks or anything, you know, that matters to you, to a model that's not even a top 20 model. Not not a good call.
Jordan Wilson [00:37:21]:
You shouldn't be doing it. So start to wrap up here. Is this is the code red warning a sign that ChatGPT has officially lost the AI crown? No. It's not. It's not. Here's why. ChatGPT still dominates. Like I said, 800,000,000 weekly active users.
Jordan Wilson [00:37:42]:
Also, OpenAI just came out with a new kind of state of AI in the enterprise report yesterday that we shared about our newsletter. That's why you gotta read the newsletter. Right? We don't miss anything. It shows it shows that their user message value has increased eight fold over the past year, which is insane. Also, the new ramp AI index shows that 36% of US businesses use ChatGPT Enterprise versus only 14% for Anthropic. So, yeah, I still don't think from a market share perspective, it's still ChatGPT. And that's gonna take a while. Right? Google's doing a lot, to claw into that.
Jordan Wilson [00:38:26]:
But also, it's all these I think the benchmark headlines and the social media drama have probably outpaced the actual real world usage shifts. And so far, it's just Google that's actually getting close in that department. So, no, ChatGPT hasn't lost its major crown. Although I think there's more crowns and maybe, you know, ChatGPT had five of them, you know, a year ago or thirteen months ago, they don't anymore. The new reality, I think there's three companies and they now hold three very different crowns. So Anthropic leads in coding and agentic models for now. Right? This could obviously change tomorrow, next week, next month. I think Google leads in reasoning multi-modality and distribution.
Jordan Wilson [00:39:10]:
Right? They have an unfair advantage, right, with Google AI mode, but also rolling out Gemini in search, Android, cloud, everywhere else. I mean, Google Gemini is everywhere. Right? I sneeze, and I think Google Gemini is in my Kleenex. Right? My wife just screenshotted me. It's like, oh my gosh. Like, Google Gemini is in, you know, Google Maps. It's everywhere. Right? The distribution advantage that they have is unfair, but still OpenAI, even after all that, they still lead in consumer mind share.
Jordan Wilson [00:39:40]:
The word AI is still synonymous with ChatGPT, and I don't think that is going to change anytime soon. Right? And the only time the only reason we're having this conversation, right, is because if you're listening to this myself, right, we are in a bubble. Right? I'm not saying though the AI bubble, it's gonna burst the .com. No. I'm saying we are in an AI bubble. Right? If you follow this frequently, if you're using it all day every day, if you're passionate, if you're an AI enthusiast, you're probably paying way closer attention to this than 99% of the world.
Jordan Wilson [00:40:20]:
I live in Chicago. Right? But, what I'm saying is, I'm kind of away from the big, kind of ongoing narrative. And I think that the narrative in Silicon Valley, in San Francisco, right out in California, it gets blown up a little bit. So, because it just seems like it's, you know, life and death, everything is is AI. And, yes, it's hard to succeed in business or even it's gonna become harder to have a personal life, that's thriving without using AI because it is literally everywhere. You can't use, you know, really any app without, using AI right now. But I think it's just this narrative has been a little bit overblown. But here's the other thing.
Jordan Wilson [00:41:04]:
Hot take to end it here. OpenAI is not going anywhere. Right? Yeah. The market share, but let me share some more stats here from a similar web. So they put out a recent, stat on app stickiness. Right? So even though OpenAI has lost some of their market share to Google Gemini, stickiness or the daily reliance, that an app has, ChatGPT has a 33% stickiness rate. Alright? And most other, large language models, Perplexity, you know, Copilot, Gemini. You know, Gemini's is actually pretty low at 5%, Grok, 14%, Claude, 16%.
Jordan Wilson [00:41:52]:
What does this mean? The people that are using ChatGPT are using it way more frequently than any other large language model. And it's not even close. It is stickier. Right? And what does that mean? And why am I bringing this up at the end? Ads. Yeah. I know people are gonna hate it, but another stat. Alright. From a similar web.
Jordan Wilson [00:42:15]:
So it showed the average query length, and I think this is important. Right? The average Google query length is 3.4, 3.4 words. The average AI mode, query length inside Google's AI mode is 10.4 words. Alright. In ChatGPT, it is 60 words. So what does this mean and why does this have anything to do with ChatGPT not losing its crown? Ads are coming. Yes. They reportedly got paused with this code red.
Jordan Wilson [00:42:45]:
Although right before that, the week before this, we did see instances or references to ChatGPT ads in the Android ChatGPT app leak. But ads are coming. Right? And, yeah, they may not be coming now until 2026. But these last two charts show you why OpenAI is in no trouble. The platform is stickier than anything else, and it's not even close. And the context that users share with ChatGPT is so much, it is much more personalized than anything else. Alright. So that's a wrap y'all.
Jordan Wilson [00:43:26]:
But let me just say this. It's probably a good thing that OpenAI has this code red. Right? Because like I said, one of the underlying issues that no one is talking about is the default model that so many people are using day to day. Maybe you've been using it for your business. Maybe you've been using it to launch a side project, just using it in your personal life. I don't know. And maybe you've noticed like a lot of people, hey. It seems like this ChatGPT is not as good as it once was.
Jordan Wilson [00:43:57]:
Or, oh my gosh. Like, you know, Salesforce CEO, Marc Benioff, when he, you know, is using ChatGPT every day and, you know, yeah, maybe there's some recency bias there since GPT-5 launched in August, not very good. GPT-5.1, not very good, especially when you're using the default chat model. But then when you go use the default chat model of Google Gemini, you're getting Gemini 2.0 or you're getting Gemini 3 Pro. Right? So there's a big drop off there. So it is probably in OpenAI's best interest to go through this code red to hit a pause even though I'd love to see, you know, Atlas get, you know, developed more and faster and and Pulse and even ads. I I I I maybe I'm the only one. I can't wait for JWT ads.
Jordan Wilson [00:44:44]:
Right? But maybe it's a good thing because maybe OpenAI is going to catch back up in the model game, the game that they technically created and dominated. But for now, Google's kind of taking the crown. But if I'm OpenAI, I'm not sweating it. Alright. I hope today's show was helpful and maybe a little bit entertaining, bringing you hot take Tuesday. Thank you for tuning in. If you missed anything, don't worry. It's all gonna be recapped on our website in today's daily newsletter at youreverydayai.com.
Jordan Wilson [00:45:14]:
Thanks for tuning in. Hope to see you back tomorrow and every day for more Everyday AI. Thanks y'all.