Go back to my posts from a month ago. I said even OpenAI's internal model wasn't on Opus 5.5's level, let alone Fable.
It was obvious to me from day one.
OpenAI can't ship a frontier model right now. They've drifted far from the basics of what makes a model good. Their post-training is bad and it degrades the model. All the instructions and restrictions piled on top end up warping what it can actually do.
That's why OpenAI models are great at analysis and terrible at execution. It's like that expert who talks like he knows everything about life, then the real work starts and he just falls apart.
This started with GPT-5.6 Sol and got worse with every model since. That's how their models ended up degraded.
Instead of owning it, they tuned out what people were saying and kept running campaigns to sell their models as frontier and competitive. That's nowhere near reality.
Look at Google. They quietly worked their way back into the race.
OpenAI doesn't have that depth. They're all over the place between ChatGPT, ChatGPT Work, Codex, Dot and Sora. Nothing's landing. No priorities, no focus. They want to compete on every front with no clear plan.
I've been saying this for a while. I told you even GPT-6.2 Astra wouldn't save them. It did the opposite. It showed people exactly why Claude and Gemini 4 matter, models that keep evolving and bring something new and exciting.
If anyone ever tells you OpenAI has a stronger model and safety is holding it back, don't buy it. They don't have one. I'm sure of it.
If they did they would have shipped it against Opus 5.5 or Fable 5.2, especially with Anthropic's IPO around the corner. Instead they leaned on hype about a legendary model that solved math problems, reinvented the universe, blah blah blah. But there's nothing there.
From today there's not a single reason to keep a ChatGPT subscription or stick with OpenAI models. The one exception is computer use in some cases, and that's maybe 2% of all use cases.
OpenAI fell hard today and I don't think they'll recover. You'll see where this leads soon, when a tech giant acquires OpenAI. They have massive potential. What they're missing is the vision to compete.
GPT-6.2 Sol & Astra are *finally* here. We've also expanded ultrafast to Pro 200 plans for GPT-6.2 Sol, it's my go-to model for most day to day things.
And here we go…GPT-6.2 Astra and GPT-6.2 Sol are yours to make more games with.
Oh, and they’re also super efficient and really good at making slides, building websites, and tackling the work you were doing before you started making games.
Rolling out today to ChatGPT Work and Codex for Plus, Pro, Enterprise, and Edu users.
https://t.co/tjQBDQWB2n
Some of you have been asking very politely. Some of you have just been asking.
GPT-6.2 Astra and Sol are now available.
Super efficient and built to perform where it counts—the real world.
https://t.co/qr2iP0GpfX