The Lacklustre Impact of GPT-5 Leads to Whispers of GPT-6

Just a week after the rather lacklustre launch of GPT-5, OpenAI’s CEO Sam Altman was already looking ahead. He has given the world an early glimpse into GPT-6 and the focus is not raw performance or speed. Instead, the spotlight is on memory.

“People want models that remember things over time,” Altman said, during a recent event.

It might sound like a small shift, but in practice this could reshape how we interact with AI.

Why GPT-5 fell flat

The launch of GPT-5 shows why memory is so important. For all the hype leading up to it, the release was widely seen as underwhelming. Axios said it “landed with a thud.” The Verge called it “not a massive step forward.” The Financial Times described the rollout as “disappointing.”

The model is competent, especially in coding tasks, but the improvements seem incremental. There is no defining feature that made people sit up and take notice. This lack of a clear step change has fuelled scepticism about whether AI is starting to plateau, especially since it took two and a half years to get from GPT-4 to GPT-5.

GPT-6’s focus on memory feels like a response to that criticism. Rather than another marginally stronger model, it promises a qualitative change in how we experience AI.

GTP-6: A more personal AI

At the moment, most AI tools feel transactional. You open a chat, type a prompt, get your result, and close the window. The model doesn’t truly remember who you are when you return tomorrow. GPT-6 aims to break this cycle by remembering the details that matter to you. That could include your tone of voice, your writing style, your goals or the projects you are working on.

Altman is clear on why this matters: “People want memory.” Without it, conversations feel limited. With it, interactions become cumulative. The AI can pick up where you left off, rather than making you repeat the same instructions every time. For businesses, that opens up possibilities. Imagine an AI that truly remembers brand guidelines, ongoing campaigns, or client preferences, without needing to be reminded every session.

Memory over muscle

“Bigger” has been the default path for AI development. More data, more parameters, more computing. But bigger is not always better. Altman has made the case that memory is what will make AI genuinely useful long-term. An AI that can anticipate your needs, recall previous conversations and evolve with you is more powerful than one that is simply larger.

Think about the difference between a search engine and a personal assistant. A search engine gives you results on demand but forgets what you asked yesterday. A personal assistant keeps track of your schedule, your preferences and your priorities. GPT-6 is positioned to be more like the latter, which could transform how professionals use AI day-to-day, from project management to client servicing.

Privacy and trust come first

Of course, memory raises questions. How much will GPT-6 remember? How will that data be stored? Who will have access to it? Without clear answers, memory could quickly cross the line from helpful to intrusive. Altman has said that user control, transparency and privacy safeguards will be built in from the start.

That means users should be able to see what is being remembered, choose what to delete, and understand how information is stored. Encryption and secure storage will be essential. Without that trust, memory becomes a liability rather than a feature.

This is where businesses will need to be cautious. Adopting AI with memory will require a serious look at data governance and compliance, not just productivity gains.

A mindful approach

OpenAI is not just thinking about the technical side. They have been consulting with psychologists to understand how long-term interaction with AI could affect people emotionally. That might sound cautious, but it is also wise. If an AI begins to feel more like a companion, people may start to form attachments. And while that may bring benefits, it also raises questions about dependency, trust and emotional wellbeing.

By taking this step, OpenAI is signalling that GPT-6 is not just a technical project, but a social one. For leaders, this is a reminder that adopting AI is about more than just efficiency. It’s about culture, behaviour and human impact too.

Progress in checkpoints

Altman has also promised that GPT-6 will not arrive in a single, dramatic drop. Instead, it will roll out gradually, in what he calls “checkpoints”. Each stage will allow for feedback, refinement and adjustment before moving forward.

This feels like a deliberate response to GPT-5’s rocky debut. A staggered rollout gives OpenAI the chance to manage expectations and learn from user experiences. For businesses, it means there will be opportunities to experiment with early versions of GPT-6, adapt processes, and shape how memory is applied in practice.

The bigger picture

GPT-6 is not simply the next model in a series. It is a statement about what makes AI valuable. Bigger models may impress on paper, but memory is what will make them truly useful.

For brands and digital teams, this raises important questions like:

• How can memory-driven AI improve customer experience?
• How can it be used responsibly without overstepping trust?
• How can leaders prepare their organisations for AI that not only responds, but truly remembers?

The technology will move quickly. The real opportunity is for businesses that can align it with strategy, culture and customer needs. GPT-6 could be the first step towards AI that feels less like a tool and more like a partner. The challenge is to make sure that partnership is built on trust.

AI is moving fast, but not every advancement will matter for your business. At Elixirr Digital, we can help you shape an AI strategy that makes sense today and can adapt to what’s next.