If you recall my previous post, you know exactly where I stand in the AI race: I’m firmly in the Gemini camp. I believe it’s been about a year since I pulled the plug on my ChatGPT subscription, and I haven’t touched any of their new models since. Even though I remain a die-hard GitHub Copilot fan and follow its updates religiously, I steer clear of the GPT options there, even when they are free. Until recently, I was relying on a combination of Gemini 2.5 Pro and Claude 4.5 for my coding workflows. However, with the arrival of Gemini 3, my toolkit has simplified significantly—it is now the only model I use.
Why I Walked Away from OpenAI
To put it simply, OpenAI’s development roadmap over the past year just hasn’t aligned with my expectations as a regular person and a Dev. Lately, OpenAI seems to have shifted its focus toward making ChatGPT more “user-friendly” by introducing features that, frankly, feel unnecessary to me. By prioritizing these UX additions, I feel they’ve neglected the core development of the GPT model itself.
The biggest proof of this stagnation was the GPT-5 fiasco. I have no doubt many of you agree with me on this. After all the hype and promising marketing maneuvers, what they actually revealed wasn’t even a proper new model—it was essentially just a model router. That was the moment I completely lost interest in the OpenAI ecosystem.
I feel similarly about Grok. While they caught some serious momentum with Grok 3—and I still believe it is the best model for real-time data—their attempts to launch new models without substantive development improvements caused me to lose faith in their direction as well.
The Google Advantage
My hype for the Google ecosystem stems from their relentless focus on performance and efficiency, and the fact that they—unlike OpenAI, Anthropic, or Grok—are not dependent on external sources for processing power. To me, Google is effectively the “owner of the internet.” It is impossible to even estimate the scale of the data they hold.
With OpenAI or Anthropic, we can make educated guesses about what they use to train their models. With Google, the scope is unfathomable. That is why I track every model Google releases with great excitement and try to be the first to test them.
Another major factor is Google AI Studio. I’ve been using it since the Gemini 1.5 era. Not only can I experience every new model for free as soon as it drops, but as a developer, having free API access to integrate these Gemini models into my applications is an absolute game-changer. I am fully aware of what a massive privilege this is.
What “Code Red” Really Means for Us
It’s not hard to guess what kind of advantages “Code Red” will bring to the table. OpenAI declared this emergency status right after Gemini 3 Pro dropped, and honestly, hearing that made me happy. It signals that OpenAI is making an immediate return to the performance and efficiency-focused development policy that defined the GPT-3.5 and GPT-4o eras.
This shift means the effort spent on polishing the ChatGPT platform interface will likely be minimized in favor of doing substantial work on model development. Because of this, I expect all eyes will turn back to OpenAI over the next six months. However, I don’t expect Google to simply hand over the baton in this race; they won’t give up their lead easily.
As last words, regardless of which tech giant claims the throne in this race, I have zero doubt that the real winners here will be us, the users.