
As early as 2024, OpenAI CEO Sam Altman began talking about GPT-5's capabilities, promising with "a high degree of scientific certainty" that it'll be smarter than GPT-4 (which he'd previously indicated that it "kind of sucks").
The executive further claimed GPT-5 might work like a virtual brain, sporting deeper thinking capabilities. He even claimed that the model might be smarter than he is.
Elsewhere, in a recent AMA session (ask me anything), OpenAI's Vice President of Research, Jerry Tworek, indicated:
"GPT-5 is our next foundational model that is meant to just make everything our models can currently do better and with less model switching."
As you may know, OpenAI retired GPT-4 from ChatGPT on April 20, 2025, replacing it with GPT-4o. And as it now seems, GPT-5 is set to join the fold soon.
Features and improvements expected to ship GPT-5

At the beginning of August, OpenAI CEO Sam Altman indicated that there a ton of new models, products, features, and more on the way (via Tom's Guide).
we have a ton of stuff to launch over the next couple of months--new models, products, features, and more.please bear with us through some probable hiccups and capacity crunches. although it may be slightly choppy, we think you'll really love what we've created for you!August 2, 2025
- Sora 2 - As you may know, OpenAI released Sora, a text-to-video AI model. But recent leaks and rumours suggest that GPT-5's long-anticipated launch could also be accompanied by the second iteration of Sora AI, stretching its capabilities to 20-second-long high-quality videos.
- Better memory - In a recent leak suggesting that GPT-5 is about to launch, the AI is touted to ship with advanced memory and better reasoning and coding capabilities.
- Enhanced agentic advances - OpenAI has been ramping up its efforts to ship agentic AI tools like ChatGPT agent. Following GPT-5's much-anticipated launch, the AI model is expected to bolster their capabilities even further, making it easier for ChatGPT to do tasks on your behalf using connectors.
- Improved coding - Over the past few years, OpenAI has shipped sophisticated AI models, and as has always been the case, these models keep improving their coding capabilities, raising concerns that these models might eventually replace programmers. GPT-5 is expected to boast exemplary coding capabilities.
- More conversational - As AI advances and scales greater heights, there's a rise in the number of conversational models emerging, featuring human-like traits that make the conversation seem more natural. This is in the tails of companies like Microsoft attempting to make Copilot a "real friend" and companion that can age.
GPT's rumoured launch date

Over the past few months, OpenAI hinted that it was going to launch GPT-5 soon. But recently, the company narrowed down the vague wait gap and indicated that it would imminently launch the much-anticipated AI model in August.
GPT-5 will ship in four different iterations:
- gpt-5: Designed for logic and multi-step tasks.
- gpt-5-mini: A lightweight version for cost-sensitive applications.
- gpt-5-nano: Optimized for speed and ideal for applications requiring low latency.
- gpt-5-chat: Designed for advanced, natural, multimodal, and context-aware conversations for enterprise applications.
OpenAI is reportedly getting ready to launch GPT-5 later today. This news comes after a now-deleted GitHub blog post alluded as much. The AI firm is expected to announced the model's launch later today during a livestream at 10am PT / 1pm ET / 6pm BST.