GPT-5 & ChatGPT Plus: Is The Subscription Still Worth It?

8/10/2025

Are We Being Robbed? A Plus Subscriber's Honest Take on GPT-5's Value

What’s up, everyone? So, the day finally came. GPT-5 is here, & honestly, the dust is still settling. As a long-time ChatGPT Plus subscriber, I’ve been riding the waves of OpenAI’s updates for a while now, & this one… this one feels different. The hype was REAL. We were promised the moon, & in some ways, OpenAI delivered a spaceship. But as I’m sitting here, looking at my $20/month subscription, I can’t help but ask the question that’s probably on a lot of our minds: are we, the loyal Plus users, getting the short end of the stick? Are we being robbed in plain sight?

Let's get into it.

The Big Bang: What Exactly IS GPT-5?

First, let's just acknowledge that GPT-5 is a BEAST of a model. OpenAI officially dropped it on August 7, 2025, & it’s not just a slightly better version of GPT-4. It’s a whole new architecture. The headline feature is this "unified system" that’s supposed to be smarter about how it answers your prompts. It has a fast, everyday model for most things, but it also has a "deeper reasoning model" called "GPT-5 thinking" for when you need it to, well, think harder. The cool part is, you don’t have to switch between them. A "real-time router" figures out which one to use based on your conversation. So, if you say something like "think hard about this," it knows to kick into high gear.

On top of that, we're seeing some pretty wild improvements across the board:

Coding is a big one. It’s gotten way better at handling complex coding tasks, especially with front-end development & debugging big chunks of code. Some early testers are saying it has a much better eye for design, like spacing & typography. It's even being rolled out in public preview for GitHub Copilot.
Creative writing has more "rhythm." That's the word on the street. It’s supposed to be better at things with "structural ambiguity," like poetry that doesn't rhyme.
It’s supposedly more accurate. OpenAI is claiming that GPT-5 has significantly fewer hallucinations & is better at following instructions. They're saying it's 80% less likely to have a factual error than o3. That's a big deal, especially for anyone using it for research or work.
Health-related answers are more reliable. This is interesting. GPT-5 is designed to be more of an "active thought partner," flagging potential concerns & asking clarifying questions. It scored way higher on OpenAI's HealthBench evaluation.
It’s multimodal. This was expected, but it's still impressive. It can handle text, images, & voice, making it a more versatile tool.

So yeah, on paper, GPT-5 is a monster. It’s faster, smarter, & more capable than anything we’ve had before. But here's where things get a little… complicated.

The New World Order: Subscription Tiers & The Plus Dilemma

Here’s the thing. With great power comes a new, more confusing subscription structure. It used to be simple: free users got the basic model, & Plus users got the best of the best. Now, it’s a bit more… tiered.

Free Users: Everyone gets a taste of GPT-5 now, which is pretty cool. But there’s a catch. Free users have a pretty tight message limit – something like 10 messages every hour for the standard model & one message a day for the "Thinking" model. After that, they get bumped down to a "mini" version.
Plus Subscribers ($20/month): This is where most of us are. We get "significantly higher usage" than free users. At launch, it was 80 messages every 3 hours for the standard model, but they've temporarily doubled that to 160 messages every 3 hours. We also get 200 messages a week for the "Thinking" model & a bigger 32k context window. We can also manually select the GPT-5 Thinking model.
Pro Subscribers ($200/month): And here's the new top dog. Pro users get unlimited access to GPT-5, plus access to something called "GPT-5 pro," which has "extended reasoning" for even tougher tasks.

You see the problem here, right? As a Plus subscriber, I'm no longer at the top of the food chain. I'm in this weird middle ground. I'm paying for more access, but I'm not getting the best access. The truly groundbreaking stuff, the unlimited power, is now locked behind a $200/month paywall. It feels like the value proposition for the Plus subscription has been… diluted. One article even called it "watered down".

And I'm not the only one feeling this way. A quick scroll through the OpenAI subreddit shows a lot of frustrated Plus users. People are hitting their message caps faster than ever & feeling like they've been hit with a "stealth nerf." Some are even talking about canceling their subscriptions & looking at competitors. The removal of older models that people were used to juggling to avoid caps has also been a sore spot.

So, Is It Worth It? A Practical Look at GPT-5 for a Plus User

Okay, so the subscription changes are a bit of a bummer. But what about the actual, day-to-day experience? Does the power of GPT-5 make up for the new limitations?

From what I’ve seen & tested, the answer is… maybe.

The improvements in coding are undeniable. I've thrown some pretty messy code at it, & it's been surprisingly good at not just fixing it, but explaining why it's fixing it that way. For anyone who codes for a living, this could be a game-changer. It’s like having a senior developer on call 24/7. One review even called it a "steady collaborator" rather than just a "ChatGPT that codes".

The "GPT-5 thinking" feature is also pretty impressive when it works well. I've used it for brainstorming complex content strategies & outlining long-form articles (like this one!), & the depth of the responses is noticeably better than GPT-4. It feels like it's taking a step back & considering the bigger picture, which is something I've always wanted from an AI assistant. The ability to handle long documents with its expanded context window is also a huge plus for research-heavy tasks.

But here's the rub. That "thinking" mode chews through your message limits. And when you hit that cap, you're back to the "mini" version, which feels like a big step down. It creates this weird incentive to ration your best prompts, which kind of defeats the purpose of having a powerful AI assistant in the first place.

And this is where the conversation about value really comes into focus. For businesses that rely on this kind of AI-powered assistance, the new structure makes a lot of sense from OpenAI's perspective. It pushes them towards the Pro subscription. But for individual creators, developers, & power users who have been happily paying our $20/month, it feels like we're being squeezed.

The Business Angle: Where Does This Leave AI Automation?

Now, let's put on our business hats for a second. The advancements in GPT-5 have some pretty exciting implications for things like customer service & website engagement. The improved reasoning & reduced hallucinations mean that AI chatbots can be more reliable & helpful than ever before.

This is where a platform like Arsturn comes into play. Arsturn helps businesses create custom AI chatbots trained on their own data. Imagine feeding all of your company's support documents, product information, & past customer conversations into a GPT-5 powered chatbot. You could have a customer service agent that's available 24/7, can answer complex questions instantly, & never gets tired or frustrated. That's a HUGE win for businesses of any size.

The improved coding abilities of GPT-5 also mean that creating these custom chatbots can be even easier. With a no-code platform like Arsturn, you don't even need to be a developer to build a powerful AI assistant for your website. You can train it on your data, customize its personality, & have it up & running in no time. For businesses looking to boost lead generation & provide personalized customer experiences, this is a pretty compelling proposition. A conversational AI platform like Arsturn, powered by the latest models, can help businesses build meaningful connections with their audience in a way that was never possible before.

The Competition: Is the Grass Greener Elsewhere?

With all this talk about value, it's only natural to look at the competition. And the AI landscape is more crowded than ever. Anthropic's Claude 3 has been a major contender, especially for coding tasks. Some developers still swear by it, even with GPT-5 on the scene. There are even some benchmarks where Claude still holds its own, particularly in certain coding scenarios.

Then there's Google's Gemini & other players like Grok. Each of these models has its own strengths & weaknesses. Gemini is known for its large context window, while Grok is praised for its real-time data access.

The point is, OpenAI is no longer the only game in town. And as a Plus subscriber, if I'm feeling like I'm not getting my money's worth, it's easier than ever to shop around. The $20/month I'm paying for ChatGPT Plus could go towards a subscription for another service that might have a better value proposition for my specific needs.

The Verdict: Robbery, or Just the Price of Progress?

So, back to the original question: are we being robbed?

Honestly, "robbed" is a strong word. We're not being held at gunpoint & forced to pay for a service we don't want. We have a choice. But does it feel like we're getting less value for our money than we used to? Absolutely.

The "forced upgrade" to GPT-5, with its new subscription tiers & usage caps, has definitely changed the game for Plus subscribers. We've gone from feeling like VIPs to feeling like we're stuck in premium economy while the real party is happening in first class.

Here’s my take on it. If you're a casual user who just wants a super-smart AI to play around with, the free tier of GPT-5 is probably more than enough for you. You'll get to experience the power of the new model without paying a dime.

If you're a business that's serious about integrating AI into your workflow, you should probably be looking at the Pro subscription or leveraging a platform like Arsturn to build custom solutions. The power & flexibility you'll get will be worth the investment.

But if you're like me – a power user, a creator, a developer who has come to rely on ChatGPT Plus as an essential part of your toolkit – then you're in a tough spot. The new GPT-5 is undeniably powerful, but the limitations of the Plus plan are a real buzzkill.

For now, I'm sticking with my Plus subscription. The temporary doubling of the message limits helps, & I'm still getting a lot of value out of the "Thinking" model. But I'm also keeping a close eye on the competition. The AI world is moving at a breakneck pace, & my loyalty is to the platform that gives me the most bang for my buck.

I'm hopeful that OpenAI will listen to the feedback from the community & maybe adjust the Plus plan to make it a little more appealing. But until then, I'll be carefully counting my messages & wondering if there's a better deal just around the corner.

I hope this was helpful. Let me know what you think. Are you sticking with your Plus subscription? Have you found a better alternative? The conversation is just getting started.