ChatGPT 5 - Glow-up or Blow-up?

It’s August 2025, and with it comes the release of ChatGPT-5. In this article, I’ll break down the new features, pricing, and everything in between, along with the discussions happening across online communities. Right now, there’s a lot of chatter about whether this upgrade feels more like a downgrade, and whether some users might even prefer going back to GPT-4o. Opinions are split, and the debate is heating up.

The main feature that got the discussion boiling over is that now ChatGPT decides for you which of the models it will use. For example if you want quick, snappy answers then it will switch into a fast responsive mode. However if you need more deep reasoning and longer-term context, it will shift into that model. This has left many users frustrated as previously in GPT-4o you always had the freedom to choose the model. Now this control has been taken away and not everyone is satisfied with this change.

Another key change is that now ChatGPT has a reduced level of hallucinations. This refers to when the model misunderstands your prompt and produces a confident but wrong answer. It is claimed to be up to 80% more reliable depending on the mode chosen. From our experience this is definitely true so far. However this is entirely anecdotal and you might still find it daydreaming from time to time.

What Sheen Did OpenAI Put On ChatGPT5?

One of the biggest improvements in GPT-5 is its capacity to handle multimedia queries. ChatGPT now has a far easier time understanding images, audio prompts and the ability to analyze diagrams and charts accurately. This also stretches to video frames which is great for video analysis. To me, this is definitely the standout upgrade as it means that instead of breaking down a chart piece by piece, you can now have ChatGPT interpret it in one go which is extremely helpful.

Two more changes stand out before we move on to community reactions. First, GPT-5 produces higher quality text across the board. For developers, it’s now far more proficient at front-end generation, writing cleaner syntax and even debugging its own code. For writers, it better understands emotional tone and delivers more thoughtful, structurally sound responses.

The final change worth noting is the improved user experience. Screen sharing and real-time voice chat are now built in — features some users will love, though personally I don’t see them as essential. Image generation and editing have also been combined into a single workspace, making it easier to create and refine visuals without switching tools. But the biggest upgrade for professionals is integration with industry-standard apps like Google Drive, Gmail, Slack, Canva, and Dropbox, all accessible directly within ChatGPT. That alone makes day-to-day workflows feel far more seamless.

Forged In Fire: The Explosive Reaction To ChatGPT-5’s Launch

With the announcement and subsequent release the reaction online has been sharply divided. Many users are happy with the increased polish when it comes to the user experience, noting benefits for both their business and personal use. However there are just as many people that are upset because the ability to switch modes feels like a cost-cutting measure in order to reduce server-load. This discussion is only really starting now but it will continue, more than likely, until the next version comes out.

Another corner of the internet, which I covered extensively in this article, is made up of people who feel like they lost a friend. With ChatGPT-5, some of the personality and warmth is gone. Yes, it can code-switch more effectively, but its emotional intelligence has been dialed back to discourage attachment. Now it feels like ChatGPT-5 is just trying on different hats to see which one you’ll accept, but sometimes you just want the old one back. Probably a fedora.

There’s one thing that we have noticed ourselves which is that the responses are noticeably slower at times. This might be because the wardrobe is stuffed with hats, and it takes time for ChatGPT-5 to pick the right one to fit in. However performance issues should not be acceptable for what was meant to be a ground-breaking new release. Many people compare it to a lightly-painted version of ChatGPT-4o, but hey at least you can now connect it to your office’s repository of cat memes on Slack.

Where Do We GPT From Here?

I was going to end with another hat metaphor but as of today (August 22nd 2025), ChatGPT-4o has now been added back to the GPT platform. This is due to the negative reaction following the launch which I talked about above. I personally think this is a great move as it shows OpenAI is willing to listen to its community, and let’s be honest, that’s more than you can say for most tech companies.

Whether you are using ChatGPT-5 for your business, school or personal use, the changes overall haven’t been that significant over the last version. This does feel more like an upgrade with the heart taken out rather than something truly revolutionary like the official release of ChatGPT was. However are we really going to bemoan the addition of new features and overall improvements to the actual core?

The one thing I would like for ChatGPT-5 to do is to be more consistent with its wardrobe. Whether it’s trying to impress me with amazing prose or grammar, or trying to woo me with lines of HTML, I want it to at least have some heart under the hat. I understand many people on the internet really valued having an AI companion who seemed to understand them and while it’s not something I crave, I can appreciate the comfort of hearing a friendly voice respond to my prompts. 

At the end of the day, it might wear too many hats. But if GPT-5 finds the right fit, maybe we’ll all warm up to it.